Ying Xiong

Kexue Fu, Xiaoyuan Luo, Linhao Qu et al.

The expensive fine-grained annotation and data scarcity have become the primary obstacles for the widespread adoption of deep learning-based Whole Slide Images (WSI) classification algorithms in clinical practice. Unlike few-shot learning methods in natural images that can leverage the labels of each image, existing few-shot WSI classification methods only utilize a small number of fine-grained labels or weakly supervised slide labels for training in order to avoid expensive fine-grained annotation. They lack sufficient mining of available WSIs, severely limiting WSI classification performance. To address the above issues, we propose a novel and efficient dual-tier few-shot learning paradigm for WSI classification, named FAST. FAST consists of a dual-level annotation strategy and a dual-branch classification framework. Firstly, to avoid expensive fine-grained annotation, we collect a very small number of WSIs at the slide level, and annotate an extremely small number of patches. Then, to fully mining the available WSIs, we use all the patches and available patch labels to build a cache branch, which utilizes the labeled patches to learn the labels of unlabeled patches and through knowledge retrieval for patch classification. In addition to the cache branch, we also construct a prior branch that includes learnable prompt vectors, using the text encoder of visual-language models for patch classification. Finally, we integrate the results from both branches to achieve WSI classification. Extensive experiments on binary and multi-class datasets demonstrate that our proposed method significantly surpasses existing few-shot classification methods and approaches the accuracy of fully supervised methods with only 0.22$\%$ annotation costs. All codes and models will be publicly available on https://github.com/fukexue/FAST.

Shangyu Wu, Ying Xiong, Yufei Cui et al.

Large language models (LLMs) have demonstrated great success in various fields, benefiting from their huge amount of parameters that store knowledge. However, LLMs still suffer from several key issues, such as hallucination problems, knowledge update issues, and lacking domain-specific expertise. The appearance of retrieval-augmented generation (RAG), which leverages an external knowledge database to augment LLMs, makes up those drawbacks of LLMs. This paper reviews all significant techniques of RAG, especially in the retriever and the retrieval fusions. Besides, tutorial codes are provided for implementing the representative techniques in RAG. This paper further discusses the RAG update, including RAG with/without knowledge update. Then, we introduce RAG evaluation and benchmarking, as well as the application of RAG in representative NLP tasks and industrial scenarios. Finally, this paper discusses RAG's future directions and challenges for promoting this field's development.

Gursimran Singh, Xinglu Wang, Yifan Hu et al.

Large Multimodal Models (LMMs) extend Large Language Models (LLMs) by handling diverse inputs such as images, audio, and video, but at the cost of adding a multimodal encoding stage that increases both computational and memory overhead. This step negatively affects key Service Level Objectives (SLOs), such as time to first token (TTFT) and time per output token (TPOT). We introduce Encode-Prefill-Decode (EPD) Disaggregation, a novel framework that separates the encoding, prefill, and decode stages onto dedicated resources. Unlike current systems, which bundle encoding and prefill together, our approach decouples these steps, unlocking new opportunities and optimizations. These include a mechanism to cache multimedia tokens for efficient transfer, a novel way to parallelize the encoding load within a request, a module for optimal resource allocation for disaggregated serving, and a novel role-switching method to handle changing workload characteristics. Experimental evaluations with popular LMMs show substantial gains in memory efficiency (up to 15x lower peak memory utilization), batch sizes (up to 22x larger), 10x more images per request, and 2.2x larger KV caches. Furthermore, it leads to significant improvements in SLO attainment (up to 90-100% improvement) and TTFT (up to 71% reduction), compared to systems that do not disaggregate. The code is available at https://github.com/vbdi/epdserve.

Wenbo Du, Lingling Han, Ying Xiong et al.

Accurate forecasting of passenger flows is critical for maintaining the efficiency and resilience of airport operations. Recent advances in patch-based Transformer models have shown strong potential in various time series forecasting tasks. However, most existing methods rely on fixed-size patch embedding, making it difficult to model the complex and heterogeneous patterns of airport passenger flows. To address this issue, this paper proposes a deformable temporal-spectral transformer named DTSFormer that integrates a multiscale deformable partitioning module and a joint temporal-spectral filtering module. Specifically, the input sequence is dynamically partitioned into multiscale temporal patches via a novel window function-based masking, enabling the extraction of heterogeneous trends across different temporal stages. Then, within each scale, a frequency-domain attention mechanism is designed to capture both high- and low-frequency components, thereby emphasizing the volatility and periodicity inherent in airport passenger flows. Finally, the resulting multi-frequency features are subsequently fused in the time domain to jointly model short-term fluctuations and long-term trends. Comprehensive experiments are conducted on real-world passenger flow data collected at Beijing Capital International Airport from January 2023 to March 2024. The results indicate that the proposed method consistently outperforms state-of-the-art forecasting models across different prediction horizons. Further analysis shows that the deformable partitioning module aligns patch lengths with dominant periods and heterogeneous trends, enabling superior capture of sudden high-frequency fluctuations.

Yixin Chen, Ying Xiong, Shangyu Wu et al.

Retrieval-augmented generation (RAG) has become a dominant paradigm for grounding large language models (LLMs) with external evidence in knowledge-intensive question answering. A core design choice is how to fuse retrieved samples into the LLMs, where existing internal fusion approaches broadly fall into query-based fusion, parametric fusion, and latent-based fusion. Despite their effectiveness at modest retrieval scales, these methods often fail to scale gracefully as the number of retrieved candidates k increases: Larger k improves evidence coverage, yet realistic top-k retrieval inevitably contains irrelevant or redundant content and increases the inference cost. To address these limitations, we propose ReFilter, a novel latent-based fusion framework that performs token-level filtering and fusion. ReFilter consists of three key components: a context encoder for encoding context features, a gated filter for weighting each token, and a token fusion module for integrating the weighted token feature into the LLM's hidden states. Our experiments across four general-domain QA benchmarks show that ReFilter consistently achieves the best average performance under both in-domain adaptation and out-of-domain transfer. ReFilter further generalizes to five biomedical QA benchmarks in zero-shot transfer without domain fine-tuning, reaching 70.01% average accuracy with Qwen2.5-14B-Instruct.

Milad Yazdani, Mahdi Mostajabdaveh, Zirui Zhou et al.

Practical deployment of Multi-Agent Systems (MAS) demands strong test-time performance, motivating methods that guide inference-time search and selectively spend compute to improve quality. We present the Multi-Agent System Process Reward Model (MASPRM). It assigns per-action, per-agent values to partial inter-agent transcripts and acts as an inference-time controller. MASPRM is trained from multi-agent Monte Carlo Tree Search (MCTS) rollouts without requiring step-level human annotations, by propagating returns to local targets. At inference, MASPRM guides step-level beam search and MCTS, focusing computation on promising branches and pruning early. On GSM8K and MATH, MASPRM-guided decoding with an outcome reward model (ORM) applied to the final answer, improves exact match (EM) over a single straight-through MAS pass by $+30.7$ and $+22.9$ points, respectively. A MASPRM trained on GSM8K transfers zero-shot to MATH without retraining, adding $8.4$ EM points at the same budget. MASPRM is a plug-in value model that estimates per-agent progress and complements verifier-style decoders, enabling more reliable, compute-aware multi-agent reasoning. Code: https://github.com/milad1378yz/MASPRM

Xiaohui Wang, Ying Xiong, Yang Wei et al.

Transformer, BERT and their variants have achieved great success in natural language processing. Since Transformer models are huge in size, serving these models is a challenge for real industrial applications. In this paper, we propose LightSeq, a highly efficient inference library for models in the Transformer family. LightSeq includes a series of GPU optimization techniques to to streamline the computation of neural layers and to reduce memory footprint. LightSeq can easily import models trained using PyTorch and Tensorflow. Experimental results on machine translation benchmarks show that LightSeq achieves up to 14x speedup compared with TensorFlow and 1.4x compared with FasterTransformer, a concurrent CUDA implementation. The code is available at https://github.com/bytedance/lightseq.

Qian Wang, Zahra Yousefijamarani, Morgan Lindsay Heisler et al.

Modern LLM applications such as deep-research assistants, coding agents, and Retrieval-Augmented Generation (RAG) systems, repeatedly process long prompt histories containing shared document or code chunks, creating significant pressure on the Key Value (KV) cache, which must operate within limited memory while sustaining high throughput and low latency. Prefix caching partially alleviates some of these costs by reusing KV cache for previously processed tokens, but limited by strict prefix matching. Position-independent caching (PIC) enables chunk-level reuse at arbitrary positions, but requires selective recomputation and positional-encoding (PE) adjustments. However, because these operations vary across queries, KV for the same chunk diverges across requests. Moreover, without page alignment, chunk KV layouts diverge in memory, preventing page sharing. These issues result in only modest HBM savings even when many requests reuse the same content. We present MEPIC, a memory-efficient PIC system that enables chunk KV reuse across positions, requests, and batches. MEPIC aligns chunk KV to paged storage, shifts recomputation from token- to block-level so only the first block is request-specific, removes positional encodings via Rotary Position Embedding (RoPE) fusion in the attention kernel, and makes remaining blocks fully shareable. These techniques eliminate most duplicate chunk KV in HBM, reducing usage by up to 2x over state-of-the-art PIC at comparable latency and accuracy, and up to 5x for long prompts, without any model changes.

Shangyu Wu, Ying Xiong, Yufei Cui et al.

Retrieval-based augmentations (RA) incorporating knowledge from an external database into language models have greatly succeeded in various knowledge-intensive (KI) tasks. However, integrating retrievals in non-knowledge-intensive (NKI) tasks is still challenging. Existing works focus on concatenating retrievals with inputs to improve model performance. Unfortunately, the use of retrieval concatenation-based augmentations causes an increase in the input length, substantially raising the computational demands of attention mechanisms. This paper proposes a new paradigm of RA named \textbf{ReFusion}, a computation-efficient Retrieval representation Fusion with bi-level optimization. Unlike previous works, ReFusion directly fuses the retrieval representations into the hidden states of models. Specifically, ReFusion leverages an adaptive retrieval integrator to seek the optimal combination of the proposed ranking schemes across different model layers. Experimental results demonstrate that the proposed ReFusion can achieve superior and robust performance in various NKI tasks.

Shangyu Wu, Hongchao Du, Ying Xiong et al.

Large Language Models (LLMs) have achieved remarkable success in natural language processing tasks, but their massive size and computational demands hinder their deployment in resource-constrained environments. Existing model pruning methods address this issue by removing redundant structures (e.g., elements, channels, layers) from the model. However, these methods employ a heuristic pruning strategy, which leads to suboptimal performance. Besides, they also ignore the data characteristics when pruning the model. To overcome these limitations, we propose EvoP, an evolutionary pruning framework for robust LLM inference. EvoP first presents a cluster-based calibration dataset sampling (CCDS) strategy for creating a more diverse calibration dataset. EvoP then introduces an evolutionary pruning pattern searching (EPPS) method to find the optimal pruning pattern. Compared to existing model pruning techniques, EvoP achieves the best performance while maintaining the best efficiency. Experiments across different LLMs and different downstream tasks validate the effectiveness of the proposed EvoP, making it a practical and scalable solution for deploying LLMs in real-world applications.

Alireza Heidari, Wei Zhang, Ying Xiong

Vector search powers transformers technology, but real-world use demands hybrid queries that combine vector similarity with attribute filters (e.g., "top document in category X, from 2023"). Current solutions trade off recall, speed, and flexibility, relying on fragile index hacks that don't scale. We introduce FusedANN (Fused Attribute-Vector Nearest Neighbor), a geometric framework that elevates filtering to ANN optimization constraints and introduces a convex fused space via a Lagrangian-like relaxation. Our method jointly embeds attributes and vectors through transformer-based convexification, turning hard filters into continuous, weighted penalties that preserve top-k semantics while enabling efficient approximate search. We prove that FusedANN reduces to exact filtering under high selectivity, gracefully relaxes to semantically nearest attributes when exact matches are insufficient, and preserves downstream ANN alpha-approximation guarantees. Empirically, FusedANN improves query throughput by eliminating brittle filtering stages, achieving superior recall-latency tradeoffs on standard hybrid benchmarks without specialized index hacks, delivering up to 3 times higher throughput and better recall than state-of-the-art hybrid and graph-based systems. Theoretically, we provide explicit error bounds and parameter selection rules that make FusedANN practical for production. This establishes a principled, scalable, and verifiable bridge between symbolic constraints and vector similarity, unlocking a new generation of filtered retrieval systems for large, hybrid, and dynamic NLP/ML workloads.

Morgan Lindsay Heisler, Linzi Xing, Ge Shi et al.

Huawei Cloud users leverage LoRA (Low-Rank Adaptation) as an efficient and scalable method to fine-tune and customize large language models (LLMs) for application-specific needs. However, tasks that require complex reasoning or deep contextual understanding are often hindered by biases or interference from the base model when using typical decoding methods like greedy or beam search. These biases can lead to generic or task-agnostic responses from the base model instead of leveraging the LoRA-specific adaptations. In this paper, we introduce Contrastive LoRA Decoding (CoLD), a novel decoding framework designed to maximize the use of task-specific knowledge in LoRA-adapted models, resulting in better downstream performance. CoLD uses contrastive decoding by scoring candidate tokens based on the divergence between the probability distributions of a LoRA-adapted expert model and the corresponding base model. This approach prioritizes tokens that better align with the LoRA's learned representations, enhancing performance for specialized tasks. While effective, a naive implementation of CoLD is computationally expensive because each decoding step requires evaluating multiple token candidates across both models. To address this, we developed an optimized kernel for Huawei's Ascend NPU. CoLD achieves up to a 5.54% increase in task accuracy while reducing end-to-end latency by 28% compared to greedy decoding. This work provides practical and efficient decoding strategies for fine-tuned LLMs in resource-constrained environments and has broad implications for applied data science in both cloud and on-premises settings.

Alireza Heidari, Amirhossein Ahmad, Wei Zhang et al.

Learned Indexes (LIs) represent a paradigm shift from traditional index structures by employing machine learning models to approximate the cumulative distribution function (CDF) of sorted data. While LIs achieve remarkable efficiency for static datasets, their performance degrades under dynamic updates: maintaining the CDF invariant (sum of F(k) equals 1) requires global model retraining, which blocks queries and limits the queries-per-second (QPS) metric. Current approaches fail to address these retraining costs effectively, rendering them unsuitable for real-world workloads with frequent updates. In this paper, we present Sig2Model, an efficient and adaptive learned index that minimizes retraining cost through three key techniques: (1) a sigmoid boosting approximation technique that dynamically adjusts the index model by approximating update-induced shifts in data distribution with localized sigmoid functions while preserving bounded error guarantees and deferring full retraining; (2) proactive update training via Gaussian mixture models (GMMs) that identifies high-update-probability regions for strategic placeholder allocation to speed up updates; and (3) a neural joint optimization framework that continuously refines both the sigmoid ensemble and GMM parameters via gradient-based learning. We evaluate Sig2Model against state-of-the-art updatable learned indexes on real-world and synthetic workloads, and show that Sig2Model reduces retraining cost by up to 20x, achieves up to 3x higher QPS, and uses up to 1000x less memory.

Yuhui Tao, Zhongwei Zhao, Zilong Wang et al.

The non-invasive assessment of increasingly incidentally discovered renal masses is a critical challenge in urologic oncology, where diagnostic uncertainty frequently leads to the overtreatment of benign or indolent tumors. In this study, we developed and validated RenalCLIP using a dataset of 27,866 CT scans from 8,809 patients across nine Chinese medical centers and the public TCIA cohort, a visual-language foundation model for characterization, diagnosis and prognosis of renal mass. The model was developed via a two-stage pre-training strategy that first enhances the image and text encoders with domain-specific knowledge before aligning them through a contrastive learning objective, to create robust representations for superior generalization and diagnostic precision. RenalCLIP achieved better performance and superior generalizability across 10 core tasks spanning the full clinical workflow of kidney cancer, including anatomical assessment, diagnostic classification, and survival prediction, compared with other state-of-the-art general-purpose CT foundation models. Especially, for complicated task like recurrence-free survival prediction in the TCIA cohort, RenalCLIP achieved a C-index of 0.726, representing a substantial improvement of approximately 20% over the leading baselines. Furthermore, RenalCLIP's pre-training imparted remarkable data efficiency; in the diagnostic classification task, it only needs 20% training data to achieve the peak performance of all baseline models even after they were fully fine-tuned on 100% of the data. Additionally, it achieved superior performance in report generation, image-text retrieval and zero-shot diagnosis tasks. Our findings establish that RenalCLIP provides a robust tool with the potential to enhance diagnostic accuracy, refine prognostic stratification, and personalize the management of patients with kidney cancer.

Zahra Yousefijamarani, Xinglu Wang, Qian Wang et al.

Modern large language model (LLM) serving systems face challenges from highly variable requests with diverse lengths, priorities, and stage-specific service-level objectives (SLOs). Meeting these requires real-time scheduling, rapid and cost-effective scaling, and support for both collocated and disaggregated Prefill/Decode (P/D) architectures. We present HyperFlexis, a unified LLM serving system that integrates algorithmic and system-level innovations to jointly optimize scheduling and scaling under multiple SLOs. It features a multi-SLO-aware scheduler that leverages budget estimation and request prioritization to ensure proactive SLO compliance for both new and ongoing requests. The system supports prefill- and decode-stage multi-SLO scheduling for P/D-disaggregated architectures and KV cache transfers. It also enables cost-effective scaling decisions, prefill-decode instance linking during scaling, and rapid P/D role transitions. To accelerate scaling and reduce cold-start latency, a device-to-device (D2D) weight transfer mechanism is proposed that lowers weight loading overhead by up to 19.39$\times$. These optimizations allow the system to achieve up to 4.44$\times$ higher SLO attainment, 65.82% lower request latency, and cost parity with state-of-the-art baselines. The code will be released soon.

Yixin Chen, Ying Xiong, Shangyu Wu et al.

Tool-augmented large language models (LLMs) leverage external functions to extend their capabilities, but inaccurate function calls can lead to inefficiencies and increased costs.Existing methods address this challenge by fine-tuning LLMs or using demonstration-based prompting, yet they often suffer from high training overhead and fail to account for inconsistent demonstration samples, which misguide the model's invocation behavior. In this paper, we trained a behavior-aligned retriever (BAR), which provides behaviorally consistent demonstrations to help LLMs make more accurate tool-using decisions. To train the BAR, we construct a corpus including different function-calling behaviors, i.e., calling or non-calling.We use the contrastive learning framework to train the BAR with customized positive/negative pairs and a dual-negative contrastive loss, ensuring robust retrieval of behaviorally consistent examples.Experiments demonstrate that our approach significantly reduces erroneous function calls while maintaining high task performance, offering a cost-effective and efficient solution for tool-augmented LLMs.

Ying Xiong, Linjing Liu, Yufei Cui et al.

Gene expression profiling provides profound insights into molecular mechanisms, but its time-consuming and costly nature often presents significant challenges. In contrast, whole-slide hematoxylin and eosin (H&E) stained histological images are readily accessible and allow for detailed examinations of tissue structure and composition at the microscopic level. Recent advancements have utilized these histological images to predict spatially resolved gene expression profiles. However, state-of-the-art works treat gene expression prediction as a multi-output regression problem, where each gene is learned independently with its own weights, failing to capture the shared dependencies and co-expression patterns between genes. Besides, existing works can only predict gene expression values for genes seen during training, limiting their ability to generalize to new, unseen genes. To address the above limitations, this paper presents GeneQuery, which aims to solve this gene expression prediction task in a question-answering (QA) manner for better generality and flexibility. Specifically, GeneQuery takes gene-related texts as queries and whole-slide images as contexts and then predicts the queried gene expression values. With such a transformation, GeneQuery can implicitly estimate the gene distribution by introducing the gene random variable. Besides, the proposed GeneQuery consists of two architecture implementations, i.e., spot-aware GeneQuery for capturing patterns between images and gene-aware GeneQuery for capturing patterns between genes. Comprehensive experiments on spatial transcriptomics datasets show that the proposed GeneQuery outperforms existing state-of-the-art methods on known and unseen genes. More results also demonstrate that GeneQuery can potentially analyze the tissue structure.

Xiaohui Wang, Yang Wei, Ying Xiong et al.

Transformer-based neural models are used in many AI applications. Training these models is expensive, as it takes huge GPU resources and long duration. It is challenging because typical data like sentences have variable lengths, and Transformer's computation patterns are more complex than convolutional neural networks. Existing systems either only focus on model inference or optimization for only BERT-like encoder models. In this paper, we present LightSeq2, a system to accelerate training for a general family of Transformer models on GPUs. We propose a series of GPU optimization techniques tailored to the specific computation flow and memory access patterns of Transformer models. LightSeq2 supports many model architectures, including BERT (encoder-only), GPT (decoder-only), Transformer (encoder-decoder), and vision Transformer. Our experiments for a variety of models and benchmarks show that LightSeq2 is consistently faster (1.4-3.5x) than previous systems on different GPUs. In particular, it gains 308% training speedup compared with existing systems on a large public machine translation benchmark (WMT14 English-German).

Ayan Chakrabarti, Ying Xiong, Steven J. Gortler et al.

We introduce a multi-scale framework for low-level vision, where the goal is estimating physical scene values from image data---such as depth from stereo image pairs. The framework uses a dense, overlapping set of image regions at multiple scales and a "local model," such as a slanted-plane model for stereo disparity, that is expected to be valid piecewise across the visual field. Estimation is cast as optimization over a dichotomous mixture of variables, simultaneously determining which regions are inliers with respect to the local model (binary variables) and the correct co-ordinates in the local model space for each inlying region (continuous variables). When the regions are organized into a multi-scale hierarchy, optimization can occur in an efficient and parallel architecture, where distributed computational units iteratively perform calculations and share information through sparse connections between parents and children. The framework performs well on a standard benchmark for binocular stereo, and it produces a distributional scene representation that is appropriate for combining with higher-level reasoning and other low-level cues.

Ayan Chakrabarti, Ying Xiong, Baochen Sun et al.

To produce images that are suitable for display, tone-mapping is widely used in digital cameras to map linear color measurements into narrow gamuts with limited dynamic range. This introduces non-linear distortion that must be undone, through a radiometric calibration process, before computer vision systems can analyze such photographs radiometrically. This paper considers the inherent uncertainty of undoing the effects of tone-mapping. We observe that this uncertainty varies substantially across color space, making some pixels more reliable than others. We introduce a model for this uncertainty and a method for fitting it to a given camera or imaging pipeline. Once fit, the model provides for each pixel in a tone-mapped digital photograph a probability distribution over linear scene colors that could have induced it. We demonstrate how these distributions can be useful for visual inference by incorporating them into estimation algorithms for a representative set of vision tasks.

Ying Xiong, Ayan Chakrabarti, Ronen Basri et al.

We develop a framework for extracting a concise representation of the shape information available from diffuse shading in a small image patch. This produces a mid-level scene descriptor, comprised of local shape distributions that are inferred separately at every image patch across multiple scales. The framework is based on a quadratic representation of local shape that, in the absence of noise, has guarantees on recovering accurate local shape and lighting. And when noise is present, the inferred local shape distributions provide useful shape information without over-committing to any particular image explanation. These local shape distributions naturally encode the fact that some smooth diffuse regions are more informative than others, and they enable efficient and robust reconstruction of object-scale shape. Experimental results show that this approach to surface reconstruction compares well against the state-of-art on both synthetic images and captured photographs.