Bin Cui

Ling Yang, Zhilong Zhang, Yang Song et al.

Diffusion models have emerged as a powerful new family of deep generative models with record-breaking performance in many applications, including image synthesis, video generation, and molecule design. In this survey, we provide an overview of the rapidly expanding body of work on diffusion models, categorizing the research into three key areas: efficient sampling, improved likelihood estimation, and handling data with special structures. We also discuss the potential for combining diffusion models with other generative models for enhanced results. We further review the wide-ranging applications of diffusion models in fields spanning from computer vision, natural language generation, temporal data modeling, to interdisciplinary applications in other scientific disciplines. This survey aims to provide a contextualized, in-depth look at the state of diffusion models, identifying the key areas of focus and pointing to potential areas for further exploration. Github: https://github.com/YangLing0818/Diffusion-Models-Papers-Survey-Taxonomy.

Ling Yang, Zixiang Zhang, Zhilong Zhang et al.

Flow matching (FM) is a general framework for defining probability paths via Ordinary Differential Equations (ODEs) to transform between noise and data samples. Recent approaches attempt to straighten these flow trajectories to generate high-quality samples with fewer function evaluations, typically through iterative rectification methods or optimal transport solutions. In this paper, we introduce Consistency Flow Matching (Consistency-FM), a novel FM method that explicitly enforces self-consistency in the velocity field. Consistency-FM directly defines straight flows starting from different times to the same endpoint, imposing constraints on their velocity values. Additionally, we propose a multi-segment training approach for Consistency-FM to enhance expressiveness, achieving a better trade-off between sampling quality and speed. Preliminary experiments demonstrate that our Consistency-FM significantly improves training efficiency by converging 4.4x faster than consistency models and 1.7x faster than rectified flow models while achieving better generation quality. Our code is available at: https://github.com/YangLing0818/consistency_flow_matching

Huaijun Jiang, Yu Shen, Yang Li et al. · eth-zurich

Black-box optimization (BBO) has a broad range of applications, including automatic machine learning, experimental design, and database knob tuning. However, users still face challenges when applying BBO methods to their problems at hand with existing software packages in terms of applicability, performance, and efficiency. This paper presents OpenBox, an open-source BBO toolkit with improved usability. It implements user-friendly interfaces and visualization for users to define and manage their tasks. The modular design behind OpenBox facilitates its flexible deployment in existing systems. Experimental results demonstrate the effectiveness and efficiency of OpenBox over existing systems. The source code of OpenBox is available at https://github.com/PKU-DAIR/open-box.

Ling Yang, Ye Tian, Minkai Xu et al.

GNN-to-MLP distillation aims to utilize knowledge distillation (KD) to learn computationally-efficient multi-layer perceptron (student MLP) on graph data by mimicking the output representations of teacher GNN. Existing methods mainly make the MLP to mimic the GNN predictions over a few class labels. However, the class space may not be expressive enough for covering numerous diverse local graph structures, thus limiting the performance of knowledge transfer from GNN to MLP. To address this issue, we propose to learn a new powerful graph representation space by directly labeling nodes' diverse local structures for GNN-to-MLP distillation. Specifically, we propose a variant of VQ-VAE to learn a structure-aware tokenizer on graph data that can encode each node's local substructure as a discrete code. The discrete codes constitute a codebook as a new graph representation space that is able to identify different local graph structures of nodes with the corresponding code indices. Then, based on the learned codebook, we propose a new distillation target, namely soft code assignments, to directly transfer the structural knowledge of each node from GNN to MLP. The resulting framework VQGraph achieves new state-of-the-art performance on GNN-to-MLP distillation in both transductive and inductive settings across seven graph datasets. We show that VQGraph with better performance infers faster than GNNs by 828x, and also achieves accuracy improvement over GNNs and stand-alone MLPs by 3.90% and 28.05% on average, respectively. Code: https://github.com/YangLing0818/VQGraph.

Ling Yang, Zhilin Huang, Yang Song et al.

Generating images from graph-structured inputs, such as scene graphs, is uniquely challenging due to the difficulty of aligning nodes and connections in graphs with objects and their relations in images. Most existing methods address this challenge by using scene layouts, which are image-like representations of scene graphs designed to capture the coarse structures of scene images. Because scene layouts are manually crafted, the alignment with images may not be fully optimized, causing suboptimal compliance between the generated images and the original scene graphs. To tackle this issue, we propose to learn scene graph embeddings by directly optimizing their alignment with images. Specifically, we pre-train an encoder to extract both global and local information from scene graphs that are predictive of the corresponding images, relying on two loss functions: masked autoencoding loss and contrastive loss. The former trains embeddings by reconstructing randomly masked image regions, while the latter trains embeddings to discriminate between compliant and non-compliant images according to the scene graph. Given these embeddings, we build a latent diffusion model to generate images from scene graphs. The resulting method, called SGDiff, allows for the semantic manipulation of generated images by modifying scene graph nodes and connections. On the Visual Genome and COCO-Stuff datasets, we demonstrate that SGDiff outperforms state-of-the-art methods, as measured by both the Inception Score and Fréchet Inception Distance (FID) metrics. We will release our source code and trained models at https://github.com/YangLing0818/SGDiff.

Yang Li, Yu Shen, Wentao Zhang et al. · eth-zurich, microsoft-research

End-to-end AutoML has attracted intensive interests from both academia and industry which automatically searches for ML pipelines in a space induced by feature engineering, algorithm/model selection, and hyper-parameter tuning. Existing AutoML systems, however, suffer from scalability issues when applying to application domains with large, high-dimensional search spaces. We present VolcanoML, a scalable and extensible framework that facilitates systematic exploration of large AutoML search spaces. VolcanoML introduces and implements basic building blocks that decompose a large search space into smaller ones, and allows users to utilize these building blocks to compose an execution plan for the AutoML problem at hand. VolcanoML further supports a Volcano-style execution model -- akin to the one supported by modern database systems -- to execute the plan constructed. Our evaluation demonstrates that, not only does VolcanoML raise the level of expressiveness for search space decomposition in AutoML, it also leads to actual findings of decomposition strategies that are significantly more efficient than the ones employed by state-of-the-art AutoML systems such as auto-sklearn. This paper is the extended version of the initial VolcanoML paper appeared in VLDB 2021.

Yujie Wang, Siwei Chen, Longzan Luo et al.

Reinforcement Learning (RL) has become pivotal for improving model capabilities yet suffers from rollout efficiency bottlenecks due to the long-tail response length distribution. While existing works mitigate the impact of long tails via prompt-level tail scheduling, we focus on the root source of inefficiency: the distribution itself. Specifically, we characterize the long-tail distribution at a finer granularity, identifying intra-prompt long tails, and revealing that they frequently consist of ineffective verbosity. To address this, we propose a novel paradigm of active distribution shaping to shape the rollout distribution towards conciseness and certainty, thereby fundamentally resolving tail-induced overheads. We achieve this through a distribution-aware trajectory sampling mechanism, which selects trajectories from a redundant exploration space for each prompt, and an adaptive redundancy allocation scheme to maximize both shaping effectiveness and system efficiency. Experiments demonstrate significant acceleration over state-of-the-art systems by up to 1.77x without compromising model performance.

Weitong Qian, Beicheng Xu, Zhongao Xie et al.

Scientific research has traditionally been human-intensive, requiring researchers to coordinate literature, ideas, experiments, manuscripts, and review responses across long project cycles. The rise of LLM-based scientific agents creates an opportunity to automate this process. Such a system must support the full research lifecycle, maintain structured persistent memory across projects, and improve its own research procedures over time. However, existing systems either partially satisfy or fail to satisfy these requirements, leaving a gap for a unified automated scientific research system. As a result, we present AutoSci, a memory-centric agentic system for the full scientific research lifecycle. AutoSci is organized around four modules. SciMem provides schema-governed research memory, separating Long-Term Knowledge Memory for reusable scientific knowledge from Active Research Memory for project-level artifacts such as ideas, experiments, manuscripts, and reviews. SciFlow executes a five-stage lifecycle from literature understanding to rebuttal through a harness that controls state, context, verification, feedback, and orchestration. SciDAG augments difficult skills with DAG-shaped multi-agent operators and reusable stage-specific templates. SciEvolve converts feedback signals from users, experiments, reviews, and external environments into versioned updates to SciMem organization, SciFlow skills, and SciDAG templates. Together, these modules make AutoSci a persistent research environment that can execute, remember, and evolve across research projects. The code repository is available at https://github.com/skyllwt/AutoSci.

Yang Li, Huaijun Jiang, Yu Shen et al. · eth-zurich

The distributed data analytic system -- Spark is a common choice for processing massive volumes of heterogeneous data, while it is challenging to tune its parameters to achieve high performance. Recent studies try to employ auto-tuning techniques to solve this problem but suffer from three issues: limited functionality, high overhead, and inefficient search. In this paper, we present a general and efficient Spark tuning framework that can deal with the three issues simultaneously. First, we introduce a generalized tuning formulation, which can support multiple tuning goals and constraints conveniently, and a Bayesian optimization (BO) based solution to solve this generalized optimization problem. Second, to avoid high overhead from additional offline evaluations in existing methods, we propose to tune parameters along with the actual periodic executions of each job (i.e., online evaluations). To ensure safety during online job executions, we design a safe configuration acquisition method that models the safe region. Finally, three innovative techniques are leveraged to further accelerate the search process: adaptive sub-space generation, approximate gradient descent, and meta-learning method. We have implemented this framework as an independent cloud service, and applied it to the data platform in Tencent. The empirical results on both public benchmarks and large-scale production tasks demonstrate its superiority in terms of practicality, generality, and efficiency. Notably, this service saves an average of 57.00% memory cost and 34.93% CPU cost on 25K in-production tasks within 20 iterations, respectively.

Yang Li, Yu Shen, Huaijun Jiang et al. · eth-zurich, pku

With the extensive applications of machine learning models, automatic hyperparameter optimization (HPO) has become increasingly important. Motivated by the tuning behaviors of human experts, it is intuitive to leverage auxiliary knowledge from past HPO tasks to accelerate the current HPO task. In this paper, we propose TransBO, a novel two-phase transfer learning framework for HPO, which can deal with the complementary nature among source tasks and dynamics during knowledge aggregation issues simultaneously. This framework extracts and aggregates source and target knowledge jointly and adaptively, where the weights can be learned in a principled manner. The extensive experiments, including static and dynamic transfer learning settings and neural architecture search, demonstrate the superiority of TransBO over the state-of-the-arts.

Yang Li, Yu Shen, Huaijun Jiang et al. · eth-zurich, pku

The tuning of hyperparameters becomes increasingly important as machine learning (ML) models have been extensively applied in data mining applications. Among various approaches, Bayesian optimization (BO) is a successful methodology to tune hyper-parameters automatically. While traditional methods optimize each tuning task in isolation, there has been recent interest in speeding up BO by transferring knowledge across previous tasks. In this work, we introduce an automatic method to design the BO search space with the aid of tuning history from past tasks. This simple yet effective approach can be used to endow many existing BO methods with transfer learning capabilities. In addition, it enjoys the three advantages: universality, generality, and safeness. The extensive experiments show that our approach considerably boosts BO by designing a promising and compact search space instead of using the entire space, and outperforms the state-of-the-arts on a wide range of benchmarks, including machine learning and deep learning tuning tasks, and neural architecture search.

Ling Yang, Jiayi Zheng, Heyuan Wang et al.

Out-of-distribution (OOD) graph generalization are critical for many real-world applications. Existing methods neglect to discard spurious or noisy features of inputs, which are irrelevant to the label. Besides, they mainly conduct instance-level class-invariant graph learning and fail to utilize the structural class relationships between graph instances. In this work, we endeavor to address these issues in a unified framework, dubbed Individual and Structural Graph Information Bottlenecks (IS-GIB). To remove class spurious feature caused by distribution shifts, we propose Individual Graph Information Bottleneck (I-GIB) which discards irrelevant information by minimizing the mutual information between the input graph and its embeddings. To leverage the structural intra- and inter-domain correlations, we propose Structural Graph Information Bottleneck (S-GIB). Specifically for a batch of graphs with multiple domains, S-GIB first computes the pair-wise input-input, embedding-embedding, and label-label correlations. Then it minimizes the mutual information between input graph and embedding pairs while maximizing the mutual information between embedding and label pairs. The critical insight of S-GIB is to simultaneously discard spurious features and learn invariant features from a high-order perspective by maintaining class relationships under multiple distributional shifts. Notably, we unify the proposed I-GIB and S-GIB to form our complementary framework IS-GIB. Extensive experiments conducted on both node- and graph-level tasks consistently demonstrate the superior generalization ability of IS-GIB. The code is available at https://github.com/YangLing0818/GraphOOD.

Zongwei Wang, Junliang Yu, Min Gao et al.

Contrastive learning (CL) has recently gained prominence in the domain of recommender systems due to its great ability to enhance recommendation accuracy and improve model robustness. Despite its advantages, this paper identifies a vulnerability of CL-based recommender systems that they are more susceptible to poisoning attacks aiming to promote individual items. Our analysis indicates that this vulnerability is attributed to the uniform spread of representations caused by the InfoNCE loss. Furthermore, theoretical and empirical evidence shows that optimizing this loss favors smooth spectral values of representations. This finding suggests that attackers could facilitate this optimization process of CL by encouraging a more uniform distribution of spectral values, thereby enhancing the degree of representation dispersion. With these insights, we attempt to reveal a potential poisoning attack against CL-based recommender systems, which encompasses a dual-objective framework: one that induces a smoother spectral value distribution to amplify the InfoNCE loss's inherent dispersion effect, named dispersion promotion; and the other that directly elevates the visibility of target items, named rank promotion. We validate the threats of our attack model through extensive experimentation on four datasets. By shedding light on these vulnerabilities, our goal is to advance the development of more robust CL-based recommender systems. The code is available at \url{https://github.com/CoderWZW/ARLib}.

Tao Zhang, Chenglin Zhu, Yanjun Shen et al.

The adeptness of Large Language Models (LLMs) in comprehending and following natural language instructions is critical for their deployment in sophisticated real-world applications. Existing evaluations mainly focus on fragmented constraints or narrow scenarios, but they overlook the comprehensiveness and authenticity of constraints from the user's perspective. To bridge this gap, we propose CFBench, a large-scale Comprehensive Constraints Following Benchmark for LLMs, featuring 1,000 curated samples that cover more than 200 real-life scenarios and over 50 NLP tasks. CFBench meticulously compiles constraints from real-world instructions and constructs an innovative systematic framework for constraint types, which includes 10 primary categories and over 25 subcategories, and ensures each constraint is seamlessly integrated within the instructions. To make certain that the evaluation of LLM outputs aligns with user perceptions, we propose an advanced methodology that integrates multi-dimensional assessment criteria with requirement prioritization, covering various perspectives of constraints, instructions, and requirement fulfillment. Evaluating current leading LLMs on CFBench reveals substantial room for improvement in constraints following, and we further investigate influencing factors and enhancement strategies. The data and code are publicly available at https://github.com/PKU-Baichuan-MLSystemLab/CFBench

Ziyu Guo, Renrui Zhang, Longtian Qiu et al.

Contrastive Language-Image Pre-training (CLIP) has been shown to learn visual representations with great transferability, which achieves promising accuracy for zero-shot classification. To further improve its downstream performance, existing works propose additional learnable modules upon CLIP and fine-tune them by few-shot training sets. However, the resulting extra training cost and data requirement severely hinder the efficiency for model deployment and knowledge transfer. In this paper, we introduce a free-lunch enhancement method, CALIP, to boost CLIP's zero-shot performance via a parameter-free Attention module. Specifically, we guide visual and textual representations to interact with each other and explore cross-modal informative features via attention. As the pre-training has largely reduced the embedding distances between two modalities, we discard all learnable parameters in the attention and bidirectionally update the multi-modal features, enabling the whole process to be parameter-free and training-free. In this way, the images are blended with textual-aware signals and the text representations become visual-guided for better adaptive zero-shot alignment. We evaluate CALIP on various benchmarks of 14 datasets for both 2D image and 3D point cloud few-shot classification, showing consistent zero-shot performance improvement over CLIP. Based on that, we further insert a small number of linear layers in CALIP's attention module and verify our robustness under the few-shot settings, which also achieves leading performance compared to existing methods. Those extensive experiments demonstrate the superiority of our approach for efficient enhancement of CLIP.

Yanzhao Qin, Tao Zhang, Tao Zhang et al.

Large Language Models (LLMs) have become instrumental across various applications, with the customization of these models to specific scenarios becoming increasingly critical. System message, a fundamental component of LLMs, is consist of carefully crafted instructions that guide the behavior of model to meet intended goals. Despite the recognized potential of system messages to optimize AI-driven solutions, there is a notable absence of a comprehensive benchmark for evaluating how well LLMs follow system messages. To fill this gap, we introduce SysBench, a benchmark that systematically analyzes system message following ability in terms of three limitations of existing LLMs: constraint violation, instruction misjudgement and multi-turn instability. Specifically, we manually construct evaluation dataset based on six prevalent types of constraints, including 500 tailor-designed system messages and multi-turn user conversations covering various interaction relationships. Additionally, we develop a comprehensive evaluation protocol to measure model performance. Finally, we conduct extensive evaluation across various existing LLMs, measuring their ability to follow specified constraints given in system messages. The results highlight both the strengths and weaknesses of existing models, offering key insights and directions for future research. The open source library SysBench is available at https://github.com/PKU-Baichuan-MLSystemLab/SysBench.

Tianyi Bai, Yang Li, Yu Shen et al. · pku, tencent-ai

A wide spectrum of design and decision problems, including parameter tuning, A/B testing and drug design, intrinsically are instances of black-box optimization. Bayesian optimization (BO) is a powerful tool that models and optimizes such expensive "black-box" functions. However, at the beginning of optimization, vanilla Bayesian optimization methods often suffer from slow convergence issue due to inaccurate modeling based on few trials. To address this issue, researchers in the BO community propose to incorporate the spirit of transfer learning to accelerate optimization process, which could borrow strength from the past tasks (source tasks) to accelerate the current optimization problem (target task). This survey paper first summarizes transfer learning methods for Bayesian optimization from four perspectives: initial points design, search space design, surrogate model, and acquisition function. Then it highlights its methodological aspects and technical details for each approach. Finally, it showcases a wide range of applications and proposes promising future directions.

Wentao Zhang, Zeang Sheng, Mingyu Yang et al. · pku, tencent-ai

Recently, graph neural networks (GNNs) have shown prominent performance in graph representation learning by leveraging knowledge from both graph structure and node features. However, most of them have two major limitations. First, GNNs can learn higher-order structural information by stacking more layers but can not deal with large depth due to the over-smoothing issue. Second, it is not easy to apply these methods on large graphs due to the expensive computation cost and high memory usage. In this paper, we present node-adaptive feature smoothing (NAFS), a simple non-parametric method that constructs node representations without parameter learning. NAFS first extracts the features of each node with its neighbors of different hops by feature smoothing, and then adaptively combines the smoothed features. Besides, the constructed node representation can further be enhanced by the ensemble of smoothed features extracted via different smoothing strategies. We conduct experiments on four benchmark datasets on two different application scenarios: node clustering and link prediction. Remarkably, NAFS with feature ensemble outperforms the state-of-the-art GNNs on these tasks and mitigates the aforementioned two limitations of most learning-based GNN counterparts.

Jie Liu, Mengting He, Xuequn Shang et al.

Graph anomaly detection (GAD) has gained increasing attention in recent years due to its critical application in a wide range of domains, such as social networks, financial risk management, and traffic analysis. Existing GAD methods can be categorized into node and edge anomaly detection models based on the type of graph objects being detected. However, these methods typically treat node and edge anomalies as separate tasks, overlooking their associations and frequent co-occurrences in real-world graphs. As a result, they fail to leverage the complementary information provided by node and edge anomalies for mutual detection. Additionally, state-of-the-art GAD methods, such as CoLA and SL-GAD, heavily rely on negative pair sampling in contrastive learning, which incurs high computational costs, hindering their scalability to large graphs. To address these limitations, we propose a novel unified graph anomaly detection framework based on bootstrapped self-supervised learning (named BOURNE). We extract a subgraph (graph view) centered on each target node as node context and transform it into a dual hypergraph (hypergraph view) as edge context. These views are encoded using graph and hypergraph neural networks to capture the representations of nodes, edges, and their associated contexts. By swapping the context embeddings between nodes and edges and measuring the agreement in the embedding space, we enable the mutual detection of node and edge anomalies. Furthermore, BOURNE can eliminate the need for negative sampling, thereby enhancing its efficiency in handling large graphs. Extensive experiments conducted on six benchmark datasets demonstrate the superior effectiveness and efficiency of BOURNE in detecting both node and edge anomalies.

Ran Yan, Youhe Jiang, Xiaonan Nie et al.

Training large language models (LLMs) is a computationally intensive task, which is typically conducted in data centers with homogeneous high-performance GPUs. In this paper, we explore an alternative approach by deploying training computations across heterogeneous GPUs to enable better flexibility and efficiency for heterogeneous resource utilization. Toward this end, we propose a novel system, HexiScale, that can flexibly support asymmetric partition of training computations in the scope of data-, pipeline-, and tensor model parallelism. We further formalize the allocation of asymmetric partitioned training computations over a set of heterogeneous GPUs as a constrained optimization problem and propose an efficient hierarchical graph partitioning algorithm. Our approach effectively allocates training computations across heterogeneous GPUs, fully leveraging the available computational power. We compare the performance of HexiScale with state-of-the-art homogeneous and heterogeneous training systems. When training LLMs at different scales (from 7B to 30B), empirical results demonstrate that: (i) compared to state-of-the-art homogeneous baselines running over homogeneous GPUs, HexiScale achieves similar performance when running over heterogeneous GPUs with the same theoretical FLOPS; (ii) compared to state-of-the-art heterogeneous baselines running on the same heterogeneous clusters, HexiScale delivers $1.5\times$ to $2.4\times$ higher throughput.

Chunan Shi, Yilei Chen, Yilin Chen et al.

Large Language Model (LLM) inference relies on key-value (KV) caches to avoid redundant attention computation. While approximate KV cache retention techniques reduce memory usage by sacrificing model accuracy, lossless approaches instead evict KV cache blocks from GPU memory and reconstruct them on demand to preserve exact outputs. Existing lossless KV cache management systems primarily base eviction decisions on access frequency or positional heuristics, without considering how different KV cache blocks affect the execution efficiency of GPU attention kernels. In this paper, we propose AsymCache, a computation-latency-aware KV cache management system for LLM inference that explicitly aligns cache residency decisions with GPU attention kernel performance, including three key components: Multi-Segment Attention (MSA) for efficient non-contiguous KV context processing, a cache eviction policy that jointly optimizes hit rate and position-aware recomputation cost, and an adaptive chunking scheduler for high hardware utilization. Experiments show that AsymCache reduces TTFT by up to 1.90-2.03x and time-per-output-token (TPOT) by 1.62-1.71x over latest baselines, confirming the effectiveness of the method in common workloads and validating its design goal of balancing computational efficiency with cache hit rate. Moreover, the low-level design of AsymCache allows seamless integration into agent serving systems such as Continuum, where it further reduces average job latency by up to 18.1%.

Yingxia Shao, Hongzheng Li, Xizhi Gu et al.

Graph neural networks (GNNs) are a type of deep learning models that are trained on graphs and have been successfully applied in various domains. Despite the effectiveness of GNNs, it is still challenging for GNNs to efficiently scale to large graphs. As a remedy, distributed computing becomes a promising solution of training large-scale GNNs, since it is able to provide abundant computing resources. However, the dependency of graph structure increases the difficulty of achieving high-efficiency distributed GNN training, which suffers from the massive communication and workload imbalance. In recent years, many efforts have been made on distributed GNN training, and an array of training algorithms and systems have been proposed. Yet, there is a lack of systematic review on the optimization techniques for the distributed execution of GNN training. In this survey, we analyze three major challenges in distributed GNN training that are massive feature communication, the loss of model accuracy and workload imbalance. Then we introduce a new taxonomy for the optimization techniques in distributed GNN training that address the above challenges. The new taxonomy classifies existing techniques into four categories that are GNN data partition, GNN batch generation, GNN execution model, and GNN communication protocol. We carefully discuss the techniques in each category. In the end, we summarize existing distributed GNN systems for multi-GPUs, GPU-clusters and CPU-clusters, respectively, and give a discussion about the future direction on distributed GNN training.

Xinyi Zhang, Zhuo Chang, Hong Wu et al. · pku, tencent-ai

Recently using machine learning (ML) based techniques to optimize modern database management systems has attracted intensive interest from both industry and academia. With an objective to tune a specific component of a DBMS (e.g., index selection, knobs tuning), the ML-based tuning agents have shown to be able to find better configurations than experienced database administrators. However, one critical yet challenging question remains unexplored -- how to make those ML-based tuning agents work collaboratively. Existing methods do not consider the dependencies among the multiple agents, and the model used by each agent only studies the effect of changing the configurations in a single component. To tune different components for DBMS, a coordinating mechanism is needed to make the multiple agents cognizant of each other. Also, we need to decide how to allocate the limited tuning budget among the agents to maximize the performance. Such a decision is difficult to make since the distribution of the reward for each agent is unknown and non-stationary. In this paper, we study the above question and present a unified coordinating framework to efficiently utilize existing ML-based agents. First, we propose a message propagation protocol that specifies the collaboration behaviors for agents and encapsulates the global tuning messages in each agent's model. Second, we combine Thompson Sampling, a well-studied reinforcement learning algorithm with a memory buffer so that our framework can allocate budget judiciously in a non-stationary environment. Our framework defines the interfaces adapted to a broad class of ML-based tuning agents, yet simple enough for integration with existing implementations and future extensions. We show that it can effectively utilize different ML-based agents and find better configurations with 1.4~14.1X speedups on the workload execution time compared with baselines.

Xinyi Gao, Wentao Zhang, Yingxia Shao et al.

Graph neural networks (GNNs) have demonstrated excellent performance in a wide range of applications. However, the enormous size of large-scale graphs hinders their applications under real-time inference scenarios. Although existing scalable GNNs leverage linear propagation to preprocess the features and accelerate the training and inference procedure, these methods still suffer from scalability issues when making inferences on unseen nodes, as the feature preprocessing requires the graph is known and fixed. To speed up the inference in the inductive setting, we propose a novel adaptive propagation order approach that generates the personalized propagation order for each node based on its topological information. This could successfully avoid the redundant computation of feature propagation. Moreover, the trade-off between accuracy and inference latency can be flexibly controlled by simple hyper-parameters to match different latency constraints of application scenarios. To compensate for the potential inference accuracy loss, we further propose Inception Distillation to exploit the multi scale reception information and improve the inference performance. Extensive experiments are conducted on four public datasets with different scales and characteristics, and the experimental results show that our proposed inference acceleration framework outperforms the SOTA graph inference acceleration baselines in terms of both accuracy and efficiency. In particular, the advantage of our proposed method is more significant on larger-scale datasets, and our framework achieves $75\times$ inference speedup on the largest Ogbn-products dataset.

Wentao Zhang, Ziqi Yin, Zeang Sheng et al.

Graph neural networks (GNNs) have achieved great success in many graph-based applications. However, the enormous size and high sparsity level of graphs hinder their applications under industrial scenarios. Although some scalable GNNs are proposed for large-scale graphs, they adopt a fixed $K$-hop neighborhood for each node, thus facing the over-smoothing issue when adopting large propagation depths for nodes within sparse regions. To tackle the above issue, we propose a new GNN architecture -- Graph Attention Multi-Layer Perceptron (GAMLP), which can capture the underlying correlations between different scales of graph knowledge. We have deployed GAMLP in Tencent with the Angel platform, and we further evaluate GAMLP on both real-world datasets and large-scale industrial datasets. Extensive experiments on these 14 graph datasets demonstrate that GAMLP achieves state-of-the-art performance while enjoying high scalability and efficiency. Specifically, it outperforms GAT by 1.3\% regarding predictive accuracy on our large-scale Tencent Video dataset while achieving up to $50\times$ training speedup. Besides, it ranks top-1 on both the leaderboards of the largest homogeneous and heterogeneous graph (i.e., ogbn-papers100M and ogbn-mag) of Open Graph Benchmark.

Xupeng Miao, Chunan Shi, Jiangfei Duan et al.

The high computational and memory requirements of generative large language models (LLMs) make it challenging to serve them cheaply. This paper aims to reduce the monetary cost for serving LLMs by leveraging preemptible GPU instances on modern clouds, which offer accesses to spare GPUs at a much cheaper price than regular instances but may be preempted by the cloud at any time. Serving LLMs on preemptible instances requires addressing challenges induced by frequent instance preemptions and the necessity of migrating instances to handle these preemptions. This paper presents SpotServe, the first distributed LLM serving system on preemptible instances. Several key techniques in SpotServe realize fast and reliable serving of generative LLMs on cheap preemptible instances. First, SpotServe dynamically adapts the LLM parallelization configuration for dynamic instance availability and fluctuating workload, while balancing the trade-off among the overall throughput, inference latency and monetary costs. Second, to minimize the cost of migrating instances for dynamic reparallelization, the task of migrating instances is formulated as a bipartite graph matching problem, which uses the Kuhn-Munkres algorithm to identify an optimal migration plan that minimizes communications. Finally, to take advantage of the grace period offered by modern clouds, we introduce stateful inference recovery, a new inference mechanism that commits inference progress at a much finer granularity and allows SpotServe to cheaply resume inference upon preemption. We evaluate on real spot instance preemption traces and various popular LLMs and show that SpotServe can reduce the P99 tail latency by 2.4 - 9.1x compared with the best existing LLM serving systems. We also show that SpotServe can leverage the price advantage of preemptive instances, saving 54% monetary cost compared with only using on-demand instances.

Xupeng Miao, Yujie Wang, Youhe Jiang et al.

Transformer models have achieved state-of-the-art performance on various domains of applications and gradually becomes the foundations of the advanced large deep learning (DL) models. However, how to train these models over multiple GPUs efficiently is still challenging due to a large number of parallelism choices. Existing DL systems either rely on manual efforts to make distributed training plans or apply parallelism combinations within a very limited search space. In this approach, we propose Galvatron, a new system framework that incorporates multiple popular parallelism dimensions and automatically finds the most efficient hybrid parallelism strategy. To better explore such a rarely huge search space, we 1) involve a decision tree to make decomposition and pruning based on some reasonable intuitions, and then 2) design a dynamic programming search algorithm to generate the optimal plan. Evaluations on four representative Transformer workloads show that Galvatron could perform automatically distributed training with different GPU memory budgets. Among all evluated scenarios, Galvatron always achieves superior system throughput compared to previous work with limited parallelism.

Yuezihan Jiang, Yu Cheng, Hanyu Zhao et al.

We introduce ZOOMER, a system deployed at Taobao, the largest e-commerce platform in China, for training and serving GNN-based recommendations over web-scale graphs. ZOOMER is designed for tackling two challenges presented by the massive user data at Taobao: low training/serving efficiency due to the huge scale of the graphs, and low recommendation quality due to the information overload which distracts the recommendation model from specific user intentions. ZOOMER achieves this by introducing a key concept, Region of Interests (ROI) in GNNs for recommendations, i.e., a neighborhood region in the graph with significant relevance to a strong user intention. ZOOMER narrows the focus from the whole graph and "zooms in" on the more relevant ROIs, thereby reducing the training/serving cost and mitigating the information overload at the same time. With carefully designed mechanisms, ZOOMER identifies the interest expressed by each recommendation request, constructs an ROI subgraph by sampling with respect to the interest, and guides the GNN to reweigh different parts of the ROI towards the interest by a multi-level attention module. Deployed as a large-scale distributed system, ZOOMER supports graphs with billions of nodes for training and thousands of requests per second for serving. ZOOMER achieves up to 14x speedup when downsizing sampling scales with comparable (even better) AUC performance than baseline methods. Besides, both the offline evaluation and online A/B test demonstrate the effectiveness of ZOOMER.

Wentao Zhang, Yu Shen, Zheyu Lin et al.

Graph neural networks (GNNs) have achieved state-of-the-art performance in various graph-based tasks. However, as mainstream GNNs are designed based on the neural message passing mechanism, they do not scale well to data size and message passing steps. Although there has been an emerging interest in the design of scalable GNNs, current researches focus on specific GNN design, rather than the general design space, limiting the discovery of potential scalable GNN models. This paper proposes PasCa, a new paradigm and system that offers a principled approach to systemically construct and explore the design space for scalable GNNs, rather than studying individual designs. Through deconstructing the message passing mechanism, PasCa presents a novel Scalable Graph Neural Architecture Paradigm (SGAP), together with a general architecture design space consisting of 150k different designs. Following the paradigm, we implement an auto-search engine that can automatically search well-performing and scalable GNN architectures to balance the trade-off between multiple criteria (e.g., accuracy and efficiency) via multi-objective optimization. Empirical studies on ten benchmark datasets demonstrate that the representative instances (i.e., PasCa-V1, V2, and V3) discovered by our system achieve consistent performance among competitive baselines. Concretely, PasCa-V3 outperforms the state-of-the-art GNN method JK-Net by 0.4\% in terms of predictive accuracy on our large industry dataset while achieving up to $28.3\times$ training speedups.

Hailin Zhang, Xiaodong Ji, Yilin Chen et al.

As the field of Large Language Models (LLMs) continues to evolve, the context length in inference is steadily growing. Key-Value Cache (KVCache), the intermediate representations of tokens within LLM inference, has now become the primary memory bottleneck due to limited GPU memory. Current methods selectively determine suitable keys and values for self-attention computation in LLMs to address the issue. However, they either fall short in maintaining model quality or result in high serving latency. Drawing inspiration from advanced embedding retrieval techniques prevalent in the data management community, we consider the storage and retrieval of KVCache as a typical embedding retrieval problem. We propose PQCache, which employs Product Quantization (PQ) to manage KVCache, maintaining model quality while ensuring low serving latency. During the prefilling phase, we apply PQ to tokens' keys for each LLM layer and head. During the autoregressive decoding phase, we use PQ codes and centroids to approximately identify important preceding tokens, then fetch the corresponding key-value pairs for self-attention computation. Through meticulous design of overlapping and caching, we minimize any additional computation and communication overhead during both phases. Extensive experiments demonstrate that PQCache achieves both effectiveness and efficiency, with 4.60% score improvement over existing methods on InfiniteBench and low system latency in both prefilling and decoding.

Xiaonan Nie, Xupeng Miao, Zilong Wang et al.

With the increasing data volume, there is a trend of using large-scale pre-trained models to store the knowledge into an enormous number of model parameters. The training of these models is composed of lots of dense algebras, requiring a huge amount of hardware resources. Recently, sparsely-gated Mixture-of-Experts (MoEs) are becoming more popular and have demonstrated impressive pretraining scalability in various downstream tasks. However, such a sparse conditional computation may not be effective as expected in practical systems due to the routing imbalance and fluctuation problems. Generally, MoEs are becoming a new data analytics paradigm in the data life cycle and suffering from unique challenges at scales, complexities, and granularities never before possible. In this paper, we propose a novel DNN training framework, FlexMoE, which systematically and transparently address the inefficiency caused by dynamic dataflow. We first present an empirical analysis on the problems and opportunities of training MoE models, which motivates us to overcome the routing imbalance and fluctuation problems by a dynamic expert management and device placement mechanism. Then we introduce a novel scheduling module over the existing DNN runtime to monitor the data flow, make the scheduling plans, and dynamically adjust the model-to-hardware mapping guided by the real-time data traffic. A simple but efficient heuristic algorithm is exploited to dynamically optimize the device placement during training. We have conducted experiments on both NLP models (e.g., BERT and GPT) and vision models (e.g., Swin). And results show FlexMoE can achieve superior performance compared with existing systems on real-world workloads -- FlexMoE outperforms DeepSpeed by 1.70x on average and up to 2.10x, and outperforms FasterMoE by 1.30x on average and up to 1.45x.

Wentao Zhang, Zeang Sheng, Ziqi Yin et al.

Graph Neural Networks (GNNs) have achieved great success in various graph mining tasks.However, drastic performance degradation is always observed when a GNN is stacked with many layers. As a result, most GNNs only have shallow architectures, which limits their expressive power and exploitation of deep neighborhoods.Most recent studies attribute the performance degradation of deep GNNs to the \textit{over-smoothing} issue. In this paper, we disentangle the conventional graph convolution operation into two independent operations: \textit{Propagation} (\textbf{P}) and \textit{Transformation} (\textbf{T}).Following this, the depth of a GNN can be split into the propagation depth ($D_p$) and the transformation depth ($D_t$). Through extensive experiments, we find that the major cause for the performance degradation of deep GNNs is the \textit{model degradation} issue caused by large $D_t$ rather than the \textit{over-smoothing} issue mainly caused by large $D_p$. Further, we present \textit{Adaptive Initial Residual} (AIR), a plug-and-play module compatible with all kinds of GNN architectures, to alleviate the \textit{model degradation} issue and the \textit{over-smoothing} issue simultaneously. Experimental results on six real-world datasets demonstrate that GNNs equipped with AIR outperform most GNNs with shallow architectures owing to the benefits of both large $D_p$ and $D_t$, while the time costs associated with AIR can be ignored.

Xinyi Liu, Yujie Wang, Fangcheng Fu et al.

Expert parallelism is vital for effectively training Mixture-of-Experts (MoE) models, enabling different devices to host distinct experts, with each device processing different input data. However, during expert parallel training, dynamic routing results in significant load imbalance among experts: a handful of overloaded experts hinder overall iteration, emerging as a training bottleneck. In this paper, we introduce LAER-MoE, an efficient MoE training framework. The core of LAER-MoE is a novel parallel paradigm, Fully Sharded Expert Parallel (FSEP), which fully partitions each expert parameter by the number of devices and restores partial experts at expert granularity through All-to-All communication during training. This allows for flexible re-layout of expert parameters during training to enhance load balancing. In particular, we perform fine-grained scheduling of communication operations to minimize communication overhead. Additionally, we develop a load balancing planner to formulate re-layout strategies of experts and routing schemes for tokens during training. We perform experiments on an A100 cluster, and the results indicate that our system achieves up to 1.69x acceleration compared to the current state-of-the-art training systems. Source code available at https://github.com/PKU-DAIR/Hetu-Galvatron/tree/laer-moe.

Peichao Lai, Zhengfeng Zhang, Wentao Zhang et al. · pku

Recently, using large language models (LLMs) for data augmentation has led to considerable improvements in unsupervised sentence embedding models. However, existing methods encounter two primary challenges: limited data diversity and high data noise. Current approaches often neglect fine-grained knowledge, such as entities and quantities, leading to insufficient diversity. Besides, unsupervised data frequently lacks discriminative information, and the generated synthetic samples may introduce noise. In this paper, we propose a pipeline-based data augmentation method via LLMs and introduce the Gaussian-decayed gradient-assisted Contrastive Sentence Embedding (GCSE) model to enhance unsupervised sentence embeddings. To tackle the issue of low data diversity, our pipeline utilizes knowledge graphs (KGs) to extract entities and quantities, enabling LLMs to generate more diverse samples. To address high data noise, the GCSE model uses a Gaussian-decayed function to limit the impact of false hard negative samples, enhancing the model's discriminative capability. Experimental results show that our approach achieves state-of-the-art performance in semantic textual similarity (STS) tasks, using fewer data samples and smaller LLMs, demonstrating its efficiency and robustness across various models.

Xinyi Gao, Wentao Zhang, Junliang Yu et al.

Graph neural networks (GNNs) have exhibited exceptional efficacy in a diverse array of applications. However, the sheer size of large-scale graphs presents a significant challenge to real-time inference with GNNs. Although existing Scalable GNNs leverage linear propagation to preprocess the features and accelerate the training and inference procedure, these methods still suffer from scalability issues when making inferences on unseen nodes, as the feature preprocessing requires the graph to be known and fixed. To further accelerate Scalable GNNs inference in this inductive setting, we propose an online propagation framework and two novel node-adaptive propagation methods that can customize the optimal propagation depth for each node based on its topological information and thereby avoid redundant feature propagation. The trade-off between accuracy and latency can be flexibly managed through simple hyper-parameters to accommodate various latency constraints. Moreover, to compensate for the inference accuracy loss caused by the potential early termination of propagation, we further propose Inception Distillation to exploit the multi-scale receptive field information within graphs. The rigorous and comprehensive experimental study on public datasets with varying scales and characteristics demonstrates that the proposed inference acceleration framework outperforms existing state-of-the-art graph inference acceleration methods in terms of accuracy and efficiency. Particularly, the superiority of our approach is notable on datasets with larger scales, yielding a 75x inference speedup on the largest Ogbn-products dataset.

Fangcheng Fu, Huanran Xue, Yong Cheng et al.

Due to the rising concerns on privacy protection, how to build machine learning (ML) models over different data sources with security guarantees is gaining more popularity. Vertical federated learning (VFL) describes such a case where ML models are built upon the private data of different participated parties that own disjoint features for the same set of instances, which fits many real-world collaborative tasks. Nevertheless, we find that existing solutions for VFL either support limited kinds of input features or suffer from potential data leakage during the federated execution. To this end, this paper aims to investigate both the functionality and security of ML modes in the VFL scenario. To be specific, we introduce BlindFL, a novel framework for VFL training and inference. First, to address the functionality of VFL models, we propose the federated source layers to unite the data from different parties. Various kinds of features can be supported efficiently by the federated source layers, including dense, sparse, numerical, and categorical features. Second, we carefully analyze the security during the federated execution and formalize the privacy requirements. Based on the analysis, we devise secure and accurate algorithm protocols, and further prove the security guarantees under the ideal-real simulation paradigm. Extensive experiments show that BlindFL supports diverse datasets and models efficiently whilst achieves robust privacy guarantees.

Yuezihan Jiang, Hao Yang, Junyang Lin et al.

Prompt Learning has recently gained great popularity in bridging the gap between pretraining tasks and various downstream tasks. It freezes Pretrained Language Models (PLMs) and only tunes a few task-related parameters (prompts) for downstream tasks, greatly reducing the cost of tuning giant models. The key enabler of this is the idea of querying PLMs with task-specific knowledge implicated in prompts. This paper reveals a major limitation of existing methods that the indiscriminate prompts for all input data in a task ignore the intrinsic knowledge from input data, resulting in sub-optimal performance. We introduce Instance-wise Prompt Tuning (IPT), the first prompt learning paradigm that injects knowledge from the input data instances to the prompts, thereby providing PLMs with richer and more concrete context information. We devise a series of strategies to produce instance-wise prompts, addressing various concerns like model quality and cost-efficiency. Across multiple tasks and resource settings, IPT significantly outperforms task-based prompt learning methods, and achieves comparable performance to conventional finetuning with only 0.5% - 1.5% of tuned parameters.

Hailin Zhang, Penghao Zhao, Xupeng Miao et al.

Learnable embedding vector is one of the most important applications in machine learning, and is widely used in various database-related domains. However, the high dimensionality of sparse data in recommendation tasks and the huge volume of corpus in retrieval-related tasks lead to a large memory consumption of the embedding table, which poses a great challenge to the training and deployment of models. Recent research has proposed various methods to compress the embeddings at the cost of a slight decrease in model quality or the introduction of other overheads. Nevertheless, the relative performance of these methods remains unclear. Existing experimental comparisons only cover a subset of these methods and focus on limited metrics. In this paper, we perform a comprehensive comparative analysis and experimental evaluation of embedding compression. We introduce a new taxonomy that categorizes these techniques based on their characteristics and methodologies, and further develop a modular benchmarking framework that integrates 14 representative methods. Under a uniform test environment, our benchmark fairly evaluates each approach, presents their strengths and weaknesses under different memory budgets, and recommends the best method based on the use case. In addition to providing useful guidelines, our study also uncovers the limitations of current methods and suggests potential directions for future research.

Fangcheng Fu, Xupeng Miao, Jiawei Jiang et al.

Vertical federated learning (VFL) is an emerging paradigm that allows different parties (e.g., organizations or enterprises) to collaboratively build machine learning models with privacy protection. In the training phase, VFL only exchanges the intermediate statistics, i.e., forward activations and backward derivatives, across parties to compute model gradients. Nevertheless, due to its geo-distributed nature, VFL training usually suffers from the low WAN bandwidth. In this paper, we introduce CELU-VFL, a novel and efficient VFL training framework that exploits the local update technique to reduce the cross-party communication rounds. CELU-VFL caches the stale statistics and reuses them to estimate model gradients without exchanging the ad hoc statistics. Significant techniques are proposed to improve the convergence performance. First, to handle the stochastic variance problem, we propose a uniform sampling strategy to fairly choose the stale statistics for local updates. Second, to harness the errors brought by the staleness, we devise an instance weighting mechanism that measures the reliability of the estimated gradients. Theoretical analysis proves that CELU-VFL achieves a similar sub-linear convergence rate as vanilla VFL training but requires much fewer communication rounds. Empirical results on both public and real-world workloads validate that CELU-VFL can be up to six times faster than the existing works.

Wentao Zhang, Yexin Wang, Zhenbang You et al.

Graph Neural Networks (GNNs) have achieved great success in various tasks, but their performance highly relies on a large number of labeled nodes, which typically requires considerable human effort. GNN-based Active Learning (AL) methods are proposed to improve the labeling efficiency by selecting the most valuable nodes to label. Existing methods assume an oracle can correctly categorize all the selected nodes and thus just focus on the node selection. However, such an exact labeling task is costly, especially when the categorization is out of the domain of individual expert (oracle). The paper goes further, presenting a soft-label approach to AL on GNNs. Our key innovations are: i) relaxed queries where a domain expert (oracle) only judges the correctness of the predicted labels (a binary question) rather than identifying the exact class (a multi-class question), and ii) new criteria of maximizing information gain propagation for active learner with relaxed queries and soft labels. Empirical studies on public datasets demonstrate that our method significantly outperforms the state-of-the-art GNN-based AL methods in terms of both accuracy and labeling cost.

Yujie Wang, Youhe Jiang, Xupeng Miao et al.

Transformer models have emerged as the leading approach for achieving state-of-the-art performance across various application domains, serving as the foundation for advanced large-scale deep learning (DL) models. However, efficiently training these models across multiple GPUs remains a complex challenge due to the abundance of parallelism options. Existing DL systems either require manual efforts to design distributed training plans or limit parallelism combinations to a constrained search space. In this paper, we present Galvatron-BMW, a novel system framework that integrates multiple prevalent parallelism dimensions and automatically identifies the most efficient hybrid parallelism strategy. To effectively navigate this vast search space, we employ a decision tree approach for decomposition and pruning based on intuitive insights. We further utilize a dynamic programming search algorithm to derive the optimal plan. Moreover, to improve resource utilization and enhance system efficiency, we propose a bi-objective optimization workflow that focuses on workload balance. Our evaluations on different Transformer models demonstrate the capabilities of Galvatron-BMW in automating distributed training under varying GPU memory constraints. Across all tested scenarios, Galvatron-BMW consistently achieves superior system throughput, surpassing previous approaches that rely on limited parallelism strategies.

Xiaonan Nie, Yi Liu, Fangcheng Fu et al.

Recent years have witnessed the unprecedented achievements of large-scale pre-trained models, especially the Transformer models. Many products and services in Tencent Inc., such as WeChat, QQ, and Tencent Advertisement, have been opted in to gain the power of pre-trained models. In this work, we present Angel-PTM, a productive deep learning system designed for pre-training and fine-tuning Transformer models. Angel-PTM can train extremely large-scale models with hierarchical memory efficiently. The key designs of Angel-PTM are the fine-grained memory management via the Page abstraction and a unified scheduling method that coordinate the computations, data movements, and communications. Furthermore, Angel-PTM supports extreme model scaling with SSD storage and implements the lock-free updating mechanism to address the SSD I/O bandwidth bottlenecks. Experimental results demonstrate that Angel-PTM outperforms existing systems by up to 114.8% in terms of maximum model scale as well as up to 88.9% in terms of training throughput. Additionally, experiments on GPT3-175B and T5-MoE-1.2T models utilizing hundreds of GPUs verify the strong scalability of Angel-PTM.

Hung Vinh Tran, Tong Chen, Hechuan Wen et al.

Content-based recommendation systems (CRSs) utilize content features to predict user-item interactions, serving as essential tools for helping users navigate information-rich web services. However, ensuring the effectiveness of CRSs requires large-scale and even continuous model training to accommodate diverse user preferences, resulting in significant computational costs and resource demands. A promising approach to this challenge is coreset selection, which identifies a small but representative subset of data samples that preserves model quality while reducing training overhead. Yet, the selected coreset is vulnerable to the pervasive noise in user-item interactions, particularly when it is minimally sized. To this end, we propose Noise-aware Coreset Selection (NaCS), a specialized framework for CRSs. NaCS constructs coresets through submodular optimization based on training gradients, while simultaneously correcting noisy labels using a progressively trained model. Meanwhile, we refine the selected coreset by filtering out low-confidence samples through uncertainty quantification, thereby avoid training with unreliable interactions. Through extensive experiments, we show that NaCS produces higher-quality coresets for CRSs while achieving better efficiency than existing coreset selection techniques. Notably, NaCS recovers 93-95\% of full-dataset training performance using merely 1\% of the training data. The source code is available at \href{https://github.com/chenxing1999/nacs}{https://github.com/chenxing1999/nacs}.

Zheng Liu, Honglin Lin, Chonghan Qin et al.

Chart reasoning is a critical capability for Vision Language Models (VLMs). However, the development of open-source models is severely hindered by the lack of high-quality training data. Existing datasets suffer from a dual challenge: synthetic charts are often simplistic and repetitive, while the associated QA pairs are prone to hallucinations and lack the reasoning depth required for complex tasks. To bridge this gap, we propose ChartVerse, a scalable framework designed to synthesize complex charts and reliable reasoning data from scratch. (1) To address the bottleneck of simple patterns, we first introduce Rollout Posterior Entropy (RPE), a novel metric that quantifies chart complexity. Guided by RPE, we develop complexity-aware chart coder to autonomously synthesize diverse, high-complexity charts via executable programs. (2) To guarantee reasoning rigor, we develop truth-anchored inverse QA synthesis. Diverging from standard generation, we adopt an answer-first paradigm: we extract deterministic answers directly from the source code, generate questions conditional on these anchors, and enforce strict consistency verification. To further elevate difficulty and reasoning depth, we filter samples based on model fail-rate and distill high-quality Chain-of-Thought (CoT) reasoning. We curate ChartVerse-SFT-600K and ChartVerse-RL-40K using Qwen3-VL-30B-A3B-Thinking as the teacher. Experimental results demonstrate that ChartVerse-8B achieves state-of-the-art performance, notably surpassing its teacher and rivaling the stronger Qwen3-VL-32B-Thinking.

Linzhuang Sun, Tianyu Guo, Hao Liang et al.

Recent advances in Text-to-SQL have achieved strong results in static, single-turn tasks, where models generate SQL queries from natural language questions. However, these systems fall short in real-world interactive scenarios, where user intents evolve and queries must be refined over multiple turns. In applications such as finance and business analytics, users iteratively adjust query constraints or dimensions based on intermediate results. To evaluate such dynamic capabilities, we introduce DySQL-Bench, a benchmark assessing model performance under evolving user interactions. Unlike previous manually curated datasets, DySQL-Bench is built through an automated two-stage pipeline of task synthesis and verification. Structured tree representations derived from raw database tables guide LLM-based task generation, followed by interaction-oriented filtering and expert validation. Human evaluation confirms 100% correctness of the synthesized data. We further propose a multi-turn evaluation framework simulating realistic interactions among an LLM-simulated user, the model under test, and an executable database. The model must adapt its reasoning and SQL generation as user intents change. DySQL-Bench covers 13 domains across BIRD and Spider 2 databases, totaling 1,072 tasks. Even GPT-4o attains only 58.34% overall accuracy and 23.81% on the Pass@5 metric, underscoring the benchmark's difficulty. All code and data are released at https://github.com/Aurora-slz/Real-World-SQL-Bench .

Penghao Zhao, Hailin Zhang, Qinhan Yu et al.

Advancements in model algorithms, the growth of foundational models, and access to high-quality datasets have propelled the evolution of Artificial Intelligence Generated Content (AIGC). Despite its notable successes, AIGC still faces hurdles such as updating knowledge, handling long-tail data, mitigating data leakage, and managing high training and inference costs. Retrieval-Augmented Generation (RAG) has recently emerged as a paradigm to address such challenges. In particular, RAG introduces the information retrieval process, which enhances the generation process by retrieving relevant objects from available data stores, leading to higher accuracy and better robustness. In this paper, we comprehensively review existing efforts that integrate RAG technique into AIGC scenarios. We first classify RAG foundations according to how the retriever augments the generator, distilling the fundamental abstractions of the augmentation methodologies for various retrievers and generators. This unified perspective encompasses all RAG scenarios, illuminating advancements and pivotal technologies that help with potential future progress. We also summarize additional enhancements methods for RAG, facilitating effective engineering and implementation of RAG systems. Then from another view, we survey on practical applications of RAG across different modalities and tasks, offering valuable references for researchers and practitioners. Furthermore, we introduce the benchmarks for RAG, discuss the limitations of current RAG systems, and suggest potential directions for future research. Github: https://github.com/PKU-DAIR/RAG-Survey.

Qifeng Cai, Hao Liang, Chang Xu et al.

The data-centric paradigm has become pivotal in AI, especially for Text-to-SQL, where performance is limited by scarce, simplistic, and low-diversity datasets. To address this, we propose Text2SQL-Flow, a SQL-aware data augmentation framework that generates large-scale, semantically valid, and structurally diverse Text-to-SQL pairs from minimal seed data. It operates across six augmentation dimensions and integrates an end-to-end pipeline featuring SQL execution verification, natural language question generation, chain-of-thought reasoning traces, and data classification. A modular Database Manager ensures cross-database compatibility and scalability. Using this framework, we build SQLFlow, a high-quality dataset of 89,544 annotated examples. We evaluate SQLFlow in two settings: (1) For open-source LLMs, fine-tuning on SQLFlow consistently improves performance across benchmarks under the same data budget. (2) For closed-source LLMs, we introduce a masked alignment retrieval method that treats SQLFlow as both knowledge base and training data for the retriever. This enables structure-aware example matching by modeling fine-grained alignments between questions and SQL queries. Experiments show our retrieval strategy outperforms existing methods, underscoring the value of SQLFlow's high-fidelity data and our novel technique. Our work establishes a scalable, data-centric foundation for advancing Text-to-SQL systems and highlights the critical role of high-quality structured data in modern AI.

Yujie Wang, Shenhan Zhu, Fangcheng Fu et al.

Recent foundation models are capable of handling multiple tasks and multiple data modalities with the unified base model structure and several specialized model components. However, efficient training of such multi-task (MT) multi-modal (MM) models poses significant system challenges due to the sophisticated model architecture and the heterogeneous workloads of different tasks and modalities. In this paper, we propose Spindle, a brand new training system tailored for resource-efficient and high-performance training of MT MM models via wavefront scheduling. The key idea of Spindle is to decompose the model execution into waves and address the joint optimization problem sequentially, including both heterogeneity-aware workload parallelization and dependency-driven execution scheduling. We build our system and evaluate it on various MT MM models. Experiments demonstrate the superior performance and efficiency of Spindle, with speedup ratio up to 71% compared to state-of-the-art training systems.

Yu Shen, Xinyuyang Ren, Yupeng Lu et al.

Distributed data analytic engines like Spark are common choices to process massive data in industry. However, the performance of Spark SQL highly depends on the choice of configurations, where the optimal ones vary with the executed workloads. Among various alternatives for Spark SQL tuning, Bayesian optimization (BO) is a popular framework that finds near-optimal configurations given sufficient budget, but it suffers from the re-optimization issue and is not practical in real production. When applying transfer learning to accelerate the tuning process, we notice two domain-specific challenges: 1) most previous work focus on transferring tuning history, while expert knowledge from Spark engineers is of great potential to improve the tuning performance but is not well studied so far; 2) history tasks should be carefully utilized, where using dissimilar ones lead to a deteriorated performance in production. In this paper, we present Rover, a deployed online Spark SQL tuning service for efficient and safe search on industrial workloads. To address the challenges, we propose generalized transfer learning to boost the tuning performance based on external knowledge, including expert-assisted Bayesian optimization and controlled history transfer. Experiments on public benchmarks and real-world tasks show the superiority of Rover over competitive baselines. Notably, Rover saves an average of 50.1% of the memory cost on 12k real-world Spark SQL tasks in 20 iterations, among which 76.2% of the tasks achieve a significant memory reduction of over 60%.

Ling Yang, Zhaochen Yu, Chenlin Meng et al.

Diffusion models have exhibit exceptional performance in text-to-image generation and editing. However, existing methods often face challenges when handling complex text prompts that involve multiple objects with multiple attributes and relationships. In this paper, we propose a brand new training-free text-to-image generation/editing framework, namely Recaption, Plan and Generate (RPG), harnessing the powerful chain-of-thought reasoning ability of multimodal LLMs to enhance the compositionality of text-to-image diffusion models. Our approach employs the MLLM as a global planner to decompose the process of generating complex images into multiple simpler generation tasks within subregions. We propose complementary regional diffusion to enable region-wise compositional generation. Furthermore, we integrate text-guided image generation and editing within the proposed RPG in a closed-loop fashion, thereby enhancing generalization ability. Extensive experiments demonstrate our RPG outperforms state-of-the-art text-to-image diffusion models, including DALL-E 3 and SDXL, particularly in multi-category object composition and text-image semantic alignment. Notably, our RPG framework exhibits wide compatibility with various MLLM architectures (e.g., MiniGPT-4) and diffusion backbones (e.g., ControlNet). Our code is available at: https://github.com/YangLing0818/RPG-DiffusionMaster