Jing Zhao

Jing Zhao, Heliang Zheng, Chaoyue Wang et al.

The advent of open-source AI communities has produced a cornucopia of powerful text-guided diffusion models that are trained on various datasets. While few explorations have been conducted on ensembling such models to combine their strengths. In this work, we propose a simple yet effective method called Saliency-aware Noise Blending (SNB) that can empower the fused text-guided diffusion models to achieve more controllable generation. Specifically, we experimentally find that the responses of classifier-free guidance are highly related to the saliency of generated images. Thus we propose to trust different models in their areas of expertise by blending the predicted noises of two diffusion models in a saliency-aware manner. SNB is training-free and can be completed within a DDIM sampling process. Additionally, it can automatically align the semantics of two noise spaces without requiring additional annotations such as masks. Extensive experiments show the impressive effectiveness of SNB in various applications. Project page is available at https://magicfusion.github.io/.

Chaoyue Ding, Shiliang Sun, Jing Zhao

Multimodal time series (MTS) anomaly detection is crucial for maintaining the safety and stability of working devices (e.g., water treatment system and spacecraft), whose data are characterized by multivariate time series with diverse modalities. Although recent deep learning methods show great potential in anomaly detection, they do not explicitly capture spatial-temporal relationships between univariate time series of different modalities, resulting in more false negatives and false positives. In this paper, we propose a multimodal spatial-temporal graph attention network (MST-GAT) to tackle this problem. MST-GAT first employs a multimodal graph attention network (M-GAT) and a temporal convolution network to capture the spatial-temporal correlation in multimodal time series. Specifically, M-GAT uses a multi-head attention module and two relational attention modules (i.e., intra- and inter-modal attention) to model modal correlations explicitly. Furthermore, MST-GAT optimizes the reconstruction and prediction modules simultaneously. Experimental results on four multimodal benchmarks demonstrate that MST-GAT outperforms the state-of-the-art baselines. Further analysis indicates that MST-GAT strengthens the interpretability of detected anomalies by locating the most anomalous univariate time series.

Jing Zhao, Li Sun, Qingli Li

End-to-end region-based object detectors like Sparse R-CNN usually have multiple cascade bounding box decoding stages, which refine the current predictions according to their previous results. Model parameters within each stage are independent, evolving a huge cost. In this paper, we find the general setting of decoding stages is actually redundant. By simply sharing parameters and making a recursive decoder, the detector already obtains a significant improvement. The recursive decoder can be further enhanced by positional encoding (PE) of the proposal box, which makes it aware of the exact locations and sizes of input bounding boxes, thus becoming adaptive to proposals from different stages during the recursion. Moreover, we also design centerness-based PE to distinguish the RoI feature element and dynamic convolution kernels at different positions within the bounding box. To validate the effectiveness of the proposed method, we conduct intensive ablations and build the full model on three recent mainstream region-based detectors. The RecusiveDet is able to achieve obvious performance boosts with even fewer model parameters and slightly increased computation cost. Codes are available at https://github.com/bravezzzzzz/RecursiveDet.

Yifan Wang, Jing Zhao, Junwei Bao et al.

Dialogue state tracking (DST) aims to predict the current dialogue state given the dialogue history. Existing methods generally exploit the utterances of all dialogue turns to assign value for each slot. This could lead to suboptimal results due to the information introduced from irrelevant utterances in the dialogue history, which may be useless and can even cause confusion. To address this problem, we propose LUNA, a sLot-tUrN Alignment enhanced approach. It first explicitly aligns each slot with its most relevant utterance, then further predicts the corresponding value based on this aligned utterance instead of all dialogue utterances. Furthermore, we design a slot ranking auxiliary task to learn the temporal correlation among slots which could facilitate the alignment. Comprehensive experiments are conducted on multi-domain task-oriented dialogue datasets, i.e., MultiWOZ 2.0, MultiWOZ 2.1, and MultiWOZ 2.2. The results show that LUNA achieves new state-of-the-art results on these datasets.

Tencent Hunyuan Team, Ao Liu, Botong Zhou et al. · tencent-ai

As Large Language Models (LLMs) rapidly advance, we introduce Hunyuan-TurboS, a novel large hybrid Transformer-Mamba Mixture of Experts (MoE) model. It synergistically combines Mamba's long-sequence processing efficiency with Transformer's superior contextual understanding. Hunyuan-TurboS features an adaptive long-short chain-of-thought (CoT) mechanism, dynamically switching between rapid responses for simple queries and deep "thinking" modes for complex problems, optimizing computational resources. Architecturally, this 56B activated (560B total) parameter model employs 128 layers (Mamba2, Attention, FFN) with an innovative AMF/MF block pattern. Faster Mamba2 ensures linear complexity, Grouped-Query Attention minimizes KV cache, and FFNs use an MoE structure. Pre-trained on 16T high-quality tokens, it supports a 256K context length and is the first industry-deployed large-scale Mamba model. Our comprehensive post-training strategy enhances capabilities via Supervised Fine-Tuning (3M instructions), a novel Adaptive Long-short CoT Fusion method, Multi-round Deliberation Learning for iterative improvement, and a two-stage Large-scale Reinforcement Learning process targeting STEM and general instruction-following. Evaluations show strong performance: overall top 7 rank on LMSYS Chatbot Arena with a score of 1356, outperforming leading models like Gemini-2.0-Flash-001 (1352) and o4-mini-2025-04-16 (1345). TurboS also achieves an average of 77.9% across 23 automated benchmarks. Hunyuan-TurboS balances high performance and efficiency, offering substantial capabilities at lower inference costs than many reasoning models, establishing a new paradigm for efficient large-scale pre-trained models.

Yongwei Zhou, Junwei Bao, Chaoqun Duan et al.

Machine reading comprehension (MRC) that requires discrete reasoning involving symbolic operations, e.g., addition, sorting, and counting, is a challenging task. According to this nature, semantic parsing-based methods predict interpretable but complex logical forms. However, logical form generation is nontrivial and even a little perturbation in a logical form will lead to wrong answers. To alleviate this issue, multi-predictor -based methods are proposed to directly predict different types of answers and achieve improvements. However, they ignore the utilization of symbolic operations and encounter a lack of reasoning ability and interpretability. To inherit the advantages of these two types of methods, we propose OPERA, an operation-pivoted discrete reasoning framework, where lightweight symbolic operations (compared with logical forms) as neural modules are utilized to facilitate the reasoning ability and interpretability. Specifically, operations are first selected and then softly executed to simulate the answer reasoning procedure. Extensive experiments on both DROP and RACENum datasets show the reasoning ability of OPERA. Moreover, further analysis verifies its interpretability.

Jing Zhao, Yifan Wang, Junwei Bao et al.

Transformer-based pre-trained models, such as BERT, have shown extraordinary success in achieving state-of-the-art results in many natural language processing applications. However, deploying these models can be prohibitively costly, as the standard self-attention mechanism of the Transformer suffers from quadratic computational cost in the input sequence length. To confront this, we propose FCA, a fine- and coarse-granularity hybrid self-attention that reduces the computation cost through progressively shortening the computational sequence length in self-attention. Specifically, FCA conducts an attention-based scoring strategy to determine the informativeness of tokens at each layer. Then, the informative tokens serve as the fine-granularity computing units in self-attention and the uninformative tokens are replaced with one or several clusters as the coarse-granularity computing units in self-attention. Experiments on GLUE and RACE datasets show that BERT with FCA achieves 2x reduction in FLOPs over original BERT with <1% loss in accuracy. We show that FCA offers a significantly better trade-off between accuracy and FLOPs compared to prior methods.

Zhe Zhao, Yudong Li, Cheng Hou et al.

Recently, the success of pre-training in text domain has been fully extended to vision, audio, and cross-modal scenarios. The proposed pre-training models of different modalities are showing a rising trend of homogeneity in their model structures, which brings the opportunity to implement different pre-training models within a uniform framework. In this paper, we present TencentPretrain, a toolkit supporting pre-training models of different modalities. The core feature of TencentPretrain is the modular design. The toolkit uniformly divides pre-training models into 5 components: embedding, encoder, target embedding, decoder, and target. As almost all of common modules are provided in each component, users can choose the desired modules from different components to build a complete pre-training model. The modular design enables users to efficiently reproduce existing pre-training models or build brand-new one. We test the toolkit on text, vision, and audio benchmarks and show that it can match the performance of the original implementations.

Junwei Bao, Yifan Wang, Jiangyong Ying et al.

Conventional autoregressive left-to-right (L2R) sequence generation faces two issues during decoding: limited to unidirectional target sequence modeling, and constrained on strong local dependencies. To address the aforementioned problem, we propose P$^3$LM, a probabilistically permuted prophet language model, which strengthens the modeling of bidirectional information and long token dependencies for sequence generation. Specifically, P$^3$LM learns to generate tokens in permuted order upon an order-aware transformer decoder, as well as to generate the corresponding future $N$ tokens with a multi-stream attention mechanism. Extensive experiments are conducted on the GLGE benchmark, which includes four datasets for summarization, two for question generation, one for conversational question answering, and one for dialog response generation, where P$^3$LM achieves state-of-the-art results compared with strong publicly available generative pre-training methods.

Jing Zhao, Shengjian Wu, Li Sun et al.

Without densely tiled anchor boxes or grid points in the image, sparse R-CNN achieves promising results through a set of object queries and proposal boxes updated in the cascaded training manner. However, due to the sparse nature and the one-to-one relation between the query and its attending region, it heavily depends on the self attention, which is usually inaccurate in the early training stage. Moreover, in a scene of dense objects, the object query interacts with many irrelevant ones, reducing its uniqueness and harming the performance. This paper proposes to use IoU between different boxes as a prior for the value routing in self attention. The original attention matrix multiplies the same size matrix computed from the IoU of proposal boxes, and they determine the routing scheme so that the irrelevant features can be suppressed. Furthermore, to accurately extract features for both classification and regression, we add two lightweight projection heads to provide the dynamic channel masks based on object query, and they multiply with the output from dynamic convs, making the results suitable for the two different tasks. We validate the proposed scheme on different datasets, including MS-COCO and CrowdHuman, showing that it significantly improves the performance and increases the model convergence speed.

Zhi Yang, Yadong Yan, Haitao Gan et al.

In the semi-supervised learning field, Graph Convolution Network (GCN), as a variant model of GNN, has achieved promising results for non-Euclidean data by introducing convolution into GNN. However, GCN and its variant models fail to safely use the information of risk unlabeled data, which will degrade the performance of semi-supervised learning. Therefore, we propose a Safe GCN framework (Safe-GCN) to improve the learning performance. In the Safe-GCN, we design an iterative process to label the unlabeled data. In each iteration, a GCN and its supervised version(S-GCN) are learned to find the unlabeled data with high confidence. The high-confidence unlabeled data and their pseudo labels are then added to the label set. Finally, both added unlabeled data and labeled ones are used to train a S-GCN which can achieve the safe exploration of the risk unlabeled data and enable safe use of large numbers of unlabeled data. The performance of Safe-GCN is evaluated on three well-known citation network datasets and the obtained results demonstrate the effectiveness of the proposed framework over several graph-based semi-supervised learning methods.

Yun Piao, Hongbo Min, Hang Su et al.

The rapid advancement of Large Language Models (LLMs) is catalyzing a shift towards autonomous AI Agents capable of executing complex, multi-step tasks. However, these agents remain brittle when faced with real-world exceptions, making Human-in-the-Loop (HITL) supervision essential for mission-critical applications. In this paper, we present AgentBay, a novel sandbox service designed from the ground up for hybrid interaction. AgentBay provides secure, isolated execution environments spanning Windows, Linux, Android, Web Browsers, and Code interpreters. Its core contribution is a unified session accessible via a hybrid control interface: An AI agent can interact programmatically via mainstream interfaces (MCP, Open Source SDK), while a human operator can, at any moment, seamlessly take over full manual control. This seamless intervention is enabled by Adaptive Streaming Protocol (ASP). Unlike traditional VNC/RDP, ASP is specifically engineered for this hybrid use case, delivering an ultra-low-latency, smoother user experience that remains resilient even in weak network environments. It achieves this by dynamically blending command-based and video-based streaming, adapting its encoding strategy based on network conditions and the current controller (AI or human). Our evaluation demonstrates strong results in security, performance, and task completion rates. In a benchmark of complex tasks, the AgentBay (Agent + Human) model achieved more than 48% success rate improvement. Furthermore, our ASP protocol reduces bandwidth consumption by up to 50% compared to standard RDP, and in end-to-end latency with around 5% reduction, especially under poor network conditions. We posit that AgentBay provides a foundational primitive for building the next generation of reliable, human-supervised autonomous systems.

Xin Liu, Shuhuan Wen, Jing Zhao et al.

The integration of cloud computing and edge computing is an effective way to achieve global consistent and real-time multi-robot Simultaneous Localization and Mapping (SLAM). Cloud computing effectively solves the problem of limited computing, communication and storage capacity of terminal equipment. However, limited bandwidth and extremely long communication links between terminal devices and the cloud result in serious performance degradation of multi-robot SLAM systems. To reduce the computational cost of feature tracking and improve the real-time performance of the robot, a lightweight SLAM method of optical flow tracking based on pyramid IMU prediction is proposed. On this basis, a centralized multi-robot SLAM system based on a robot-edge-cloud layered architecture is proposed to realize real-time collaborative SLAM. It avoids the problems of limited on-board computing resources and low execution efficiency of single robot. In this framework, only the feature points and keyframe descriptors are transmitted and lossless encoding and compression are carried out to realize real-time remote information transmission with limited bandwidth resources. This design reduces the actual bandwidth occupied in the process of data transmission, and does not cause the loss of SLAM accuracy caused by data compression. Through experimental verification on the EuRoC dataset, compared with the current most advanced local feature compression method, our method can achieve lower data volume feature transmission, and compared with the current advanced centralized multi-robot SLAM scheme, it can achieve the same or better positioning accuracy under low computational load.

Jing Zhao, Haoyu Wang, Jinpeng Li et al.

This paper describes the THUEE team's speech recognition system for the IARPA Open Automatic Speech Recognition Challenge (OpenASR21), with further experiment explorations. We achieve outstanding results under both the Constrained and Constrained-plus training conditions. For the Constrained training condition, we construct our basic ASR system based on the standard hybrid architecture. To alleviate the Out-Of-Vocabulary (OOV) problem, we extend the pronunciation lexicon using Grapheme-to-Phoneme (G2P) techniques for both OOV and potential new words. Standard acoustic model structures such as CNN-TDNN-F and CNN-TDNN-F-A are adopted. In addition, multiple data augmentation techniques are applied. For the Constrained-plus training condition, we use the self-supervised learning framework wav2vec2.0. We experiment with various fine-tuning techniques with the Connectionist Temporal Classification (CTC) criterion on top of the publicly available pre-trained model XLSR-53. We find that the frontend feature extractor plays an important role when applying the wav2vec2.0 pre-trained model to the encoder-decoder based CTC/Attention ASR architecture. Extra improvements can be achieved by using the CTC model finetuned in the target language as the frontend feature extractor.

Ziyuan Zhu, Keyu Hu, Zhifei Chen et al.

Reconstructing continuous physical fields from sparse measurements is a central inverse problem, but data-driven generative models can produce states that violate governing dynamics. We introduce a physics-informed generative solver that separates stable prior learning from inference-time enforcement of conservation laws. Martingale-Regularized Score Matching regularizes score pretraining with a Score Fokker-Planck constraint, yielding a dynamically stable prior. Physics-Informed Implicit Score Sampling then guides denoising trajectories by gradients of physical residuals, projecting samples toward admissible manifolds without retraining. In acoustics, the method co-generates pressure and particle velocity from sparse sensors, enabling dense virtual arrays that suppress spatial aliasing. The same framework generalizes to real-world ERA5 meteorological fields under extreme sparsity. Together, this work establishes a rigorous and generalizable paradigm for solving high-dimensional inverse problems, bridging the gap between generative artificial intelligence and first-principles science.

Zhuang Yu, Lei Shen, Jing Zhao et al.

Recent multimodal large language models (MLLMs) have shown remarkable progress across vision, audio, and language tasks, yet their performance on long-form, knowledge-intensive, and temporally structured educational content remains largely unexplored. To bridge this gap, we introduce LEMON, a Lecture-based Evaluation benchmark for MultimOdal uNderstanding, focusing on STEM lecture videos that require long-horizon reasoning and cross-modal integration. LEMON comprises 2,277 video segments spanning 5 disciplines and 29 courses, with an average duration of 196.1 seconds, yielding 4,181 high-quality QA pairs, including 3,413 multiple-choice and 768 open-ended questions. Distinct from existing video benchmarks, LEMON features: (1) semantic richness and disciplinary density, (2) tightly coupled video-audio-text modalities, (3) explicit temporal and pedagogical structure, and (4) contextually linked multi-turn questioning. It further encompasses six major tasks and twelve subtasks, covering the full cognitive spectrum from perception to reasoning and then to generation. Comprehensive experiments reveal substantial performance gaps across tasks, highlighting that even state-of-the-art MLLMs like GPT-4o struggle with temporal reasoning and instructional prediction. We expect LEMON to serve as an extensible and challenging benchmark for advancing multimodal perception, reasoning, and generation in long-form instructional contents.

Jing Zhao, KokSheik Wong, Vishnu Monn Baskaran et al.

The sonata form is a musically rich and hierarchically structured form that poses significant challenges for automatic analysis. While music structure analysis has seen strides of progress in recent years, sonata form analysis remains in its early stages. This is largely due to the time-consuming and high barrier of the music background requirement for annotating classical music structures. To advance research in this area, we curated SoSA-Moz, the first large-scale dataset featuring comprehensive hierarchical structure annotations. This work establishes a foundation for systematic sonata form analysis. Leveraging this newly contributed resource, we further propose Sonalyzer-Moz, a baseline model specifically designed for investigating complex sonata structures. This framework integrates feature aggregation with sequential modeling, enabling it to capture both local feature and upper-level structural dependencies. Experiment results show that Sonalyzer-Moz is capable of identifying the components' boundaries of the upper-level structure that are critical to understanding sonata form. Therefore, this method demonstrates, for the first time, the effectiveness of automatic upper-level analysis of sonata form, and provides a robust baseline for future research in the automatic understanding of sonata form while advancing the study of classical music structure analysis.

Junxin Lu, Tengfei Song, Zhanglin Wu et al.

Text Image Machine Translation (TIMT) aims to translate text embedded in images in the source-language into target-language, requiring synergistic integration of visual perception and linguistic understanding. Existing TIMT methods, whether cascaded pipelines or end-to-end multimodal large language models (MLLMs),struggle with high-resolution text-rich images due to cluttered layouts, diverse fonts, and non-textual distractions, resulting in text omission, semantic drift, and contextual inconsistency. To address these challenges, we propose GLoTran, a global-local dual visual perception framework for MLLM-based TIMT. GLoTran integrates a low-resolution global image with multi-scale region-level text image slices under an instruction-guided alignment strategy, conditioning MLLMs to maintain scene-level contextual consistency while faithfully capturing fine-grained textual details. Moreover, to realize this dual-perception paradigm, we construct GLoD, a large-scale text-rich TIMT dataset comprising 510K high-resolution global-local image-text pairs covering diverse real-world scenarios. Extensive experiments demonstrate that GLoTran substantially improves translation completeness and accuracy over state-of-the-art MLLMs, offering a new paradigm for fine-grained TIMT under high-resolution and text-rich conditions.

Jiwoo Park, Ruoqi Liu, Avani Jagdale et al.

We present ClinicalTrialsHub, an interactive search-focused platform that consolidates all data from ClinicalTrials.gov and augments it by automatically extracting and structuring trial-relevant information from PubMed research articles. Our system effectively increases access to structured clinical trial data by 83.8% compared to relying on ClinicalTrials.gov alone, with potential to make access easier for patients, clinicians, researchers, and policymakers, advancing evidence-based medicine. ClinicalTrialsHub uses large language models such as GPT-5.1 and Gemini-3-Pro to enhance accessibility. The platform automatically parses full-text research articles to extract structured trial information, translates user queries into structured database searches, and provides an attributed question-answering system that generates evidence-grounded answers linked to specific source sentences. We demonstrate its utility through a user study involving clinicians, clinical researchers, and PhD students of pharmaceutical sciences and nursing, and a systematic automatic evaluation of its information extraction and question answering capabilities.

Kaiqi Zhang, Jing Zhao, Rui Chen

Large Language Models (LLMs) exhibit high inference latency due to their autoregressive decoding nature. While the draft head in speculative decoding mitigates this issue, its full potential remains unexplored. In this paper, we introduce KOALA (K-layer Optimized Adversarial Learning Architecture), an orthogonal approach to the draft head. By transforming the conventional single-layer draft head into a multi-layer architecture and incorporating adversarial learning into the traditional supervised training, KOALA significantly improves the accuracy of the draft head in predicting subsequent tokens, thus more closely mirroring the functionality of LLMs. Although this improvement comes at the cost of slightly increased drafting overhead, KOALA substantially unlocks the draft head's potential, greatly enhancing speculative decoding. We conducted comprehensive evaluations of KOALA, including both autoregressive and non-autoregressive draft heads across various tasks, demonstrating a latency speedup ratio improvement of 0.24x-0.41x, which is 10.57%-14.09% faster than the original draft heads.

Kaiqi Zhang, Mingguan Yang, Dali Chang et al.

Iterative method selection is crucial for solving sparse linear systems because these methods inherently lack robustness. Though image-based selection approaches have shown promise, their feature extraction techniques might encode distinct matrices into identical image representations, leading to the same selection and suboptimal method. In this paper, we introduce RAF (Relative-Absolute Fusion), an efficient feature extraction technique to enhance image-based selection approaches. By simultaneously extracting and fusing image representations as relative features with corresponding numerical values as absolute features, RAF achieves comprehensive matrix representations that prevent feature ambiguity across distinct matrices, thus improving selection accuracy and unlocking the potential of image-based selection approaches. We conducted comprehensive evaluations of RAF on SuiteSparse and our developed BMCMat (Balanced Multi-Classification Matrix dataset), demonstrating solution time reductions of 0.08s-0.29s for sparse linear systems, which is 5.86%-11.50% faster than conventional image-based selection approaches and achieves state-of-the-art (SOTA) performance. BMCMat is available at https://github.com/zkqq/BMCMat.

Wenliang Li, Rui Yan, Xu Zhang et al.

Large language models (LLMs) have demonstrated notable potential in medical applications, yet they face substantial challenges in handling complex real-world clinical diagnoses using conventional prompting methods. Current prompt engineering and multi-agent approaches typically optimize isolated inferences, neglecting the accumulation of reusable clinical experience. To address this, this study proposes a novel Multi-Agent Clinical Diagnosis (MACD) framework, which allows LLMs to self-learn clinical knowledge via a multi-agent pipeline that summarizes, refines, and applies diagnostic insights. It mirrors how physicians develop expertise through experience, enabling more focused and accurate diagnosis on key disease-specific cues. We further extend it to a MACD-human collaborative workflow, where multiple LLM-based diagnostician agents engage in iterative consultations, supported by an evaluator agent and human oversight for cases where agreement is not reached. Evaluated on 4,390 real-world patient cases across seven diseases using diverse open-source LLMs (Llama-3.1 8B/70B, DeepSeek-R1-Distill-Llama 70B), MACD significantly improves primary diagnostic accuracy, outperforming established clinical guidelines with gains up to 22.3% (MACD). In direct comparison with physician-only diagnosis under the same evaluation protocol, MACD achieves comparable or superior performance, with improvements up to 16%. Furthermore, the MACD-human workflow yields an 18.6% improvement over physician-only diagnosis, demonstrating the synergistic potential of human-AI collaboration. Notably, the self-learned clinical knowledge exhibits strong cross-model stability, transferability across LLMs, and capacity for model-specific personalization.This work thus presents a scalable self-learning paradigm that bridges the gap between the intrinsic knowledge of LLMs.

Jie Peng, Hongwei Yang, Jing Zhao et al.

Deep neural networks are vulnerable to backdoor attacks, where malicious behaviors are implanted during training. While existing defenses can effectively purify compromised models, they typically require labeled data or specific training procedures, making them difficult to apply beyond supervised learning settings. Notably, recent studies have shown successful backdoor attacks across various learning paradigms, highlighting a critical security concern. To address this gap, we propose Two-stage Symmetry Connectivity (TSC), a novel backdoor purification defense that operates independently of data format and requires only a small fraction of clean samples. Through theoretical analysis, we prove that by leveraging permutation invariance in neural networks and quadratic mode connectivity, TSC amplifies the loss on poisoned samples while maintaining bounded clean accuracy. Experiments demonstrate that TSC achieves robust performance comparable to state-of-the-art methods in supervised learning scenarios. Furthermore, TSC generalizes to self-supervised learning frameworks, such as SimCLR and CLIP, maintaining its strong defense capabilities. Our code is available at https://github.com/JiePeng104/TSC.

Le Xiao, Yunfei Xu, Jing Zhao

Domain-specific Named Entity Recognition (NER), whose goal is to recognize domain-specific entities and their categories, provides an important support for constructing domain knowledge graphs. Currently, deep learning-based methods are widely used and effective in NER tasks, but due to the reliance on large-scale labeled data. As a result, the scarcity of labeled data in a specific domain will limit its application.Therefore, many researches started to introduce few-shot methods and achieved some results. However, the entity structures in specific domains are often complex, and the current few-shot methods are difficult to adapt to NER tasks with complex features.Taking the Chinese coal chemical industry domain as an example,there exists a complex structure of multiple entities sharing a single entity, as well as multiple relationships for the same pair of entities, which affects the NER task under the sample less condition.In this paper, we propose a Large Language Models (LLMs)-based entity recognition framework LLM-DER for the domain-specific entity recognition problem in Chinese, which enriches the entity information by generating a list of relationships containing entity types through LLMs, and designing a plausibility and consistency evaluation method to remove misrecognized entities, which can effectively solve the complex structural entity recognition problem in a specific domain.The experimental results of this paper on the Resume dataset and the self-constructed coal chemical dataset Coal show that LLM-DER performs outstandingly in domain-specific entity recognition, not only outperforming the existing GPT-3.5-turbo baseline, but also exceeding the fully-supervised baseline, verifying its effectiveness in entity recognition.

Songsong Zhang, Chuanqi Tang, Hongguang Zhang et al.

Identity-Preserving Personalized Generation (IPPG) has advanced film production and artistic creation, yet existing approaches overemphasize facial regions, resulting in outputs dominated by facial close-ups.These methods suffer from weak visual narrativity and poor semantic consistency under complex text prompts, with the core limitation rooted in identity (ID) feature embeddings undermining the semantic expressiveness of generative models. To address these issues, this paper presents an IPPG method that breaks the constraint of facial close-ups, achieving synergistic optimization of identity fidelity and scene semantic creation. Specifically, we design a Dual-Line Inference (DLI) pipeline with identity-semantic separation, resolving the representation conflict between ID and semantics inherent in traditional single-path architectures. Further, we propose an Identity Adaptive Fusion (IdAF) strategy that defers ID-semantic fusion to the noise prediction stage, integrating adaptive attention fusion and noise decision masking to avoid ID embedding interference on semantics without manual masking. Finally, an Identity Aggregation Prepending (IdAP) module is introduced to aggregate ID information and replace random initializations, further enhancing identity preservation. Experimental results validate that our method achieves stable and effective performance in IPPG tasks beyond facial close-ups, enabling efficient generation without manual masking or fine-tuning. As a plug-and-play component, it can be rapidly deployed in existing IPPG frameworks, addressing the over-reliance on facial close-ups, facilitating film-level character-scene creation, and providing richer personalized generation capabilities for related domains.

Xindi Luo, Zequn Sun, Jing Zhao et al.

Parameter-efficient finetuning (PEFT) is a key technique for adapting large language models (LLMs) to downstream tasks. In this paper, we study leveraging knowledge graph embeddings to improve the effectiveness of PEFT. We propose a knowledgeable adaptation method called KnowLA. It inserts an adaptation layer into an LLM to integrate the embeddings of entities appearing in the input text. The adaptation layer is trained in combination with LoRA on instruction data. Experiments on six benchmarks with two popular LLMs and three knowledge graphs demonstrate the effectiveness and robustness of KnowLA. We show that \modelname can help activate the relevant parameterized knowledge in an LLM to answer a question without changing its parameters or input prompts.

Jing Zhao, Heliang Zheng, Chaoyue Wang et al.

Large-scale text-to-image diffusion models, (e.g., DALL-E, SDXL) are capable of generating famous persons by simply referring to their names. Is it possible to make such models generate generic identities as simple as the famous ones, e.g., just use a name? In this paper, we explore the existence of a "Name Space", where any point in the space corresponds to a specific identity. Fortunately, we find some clues in the feature space spanned by text embedding of celebrities' names. Specifically, we first extract the embeddings of celebrities' names in the Laion5B dataset with the text encoder of diffusion models. Such embeddings are used as supervision to learn an encoder that can predict the name (actually an embedding) of a given face image. We experimentally find that such name embeddings work well in promising the generated image with good identity consistency. Note that like the names of celebrities, our predicted name embeddings are disentangled from the semantics of text inputs, making the original generation capability of text-to-image models well-preserved. Moreover, by simply plugging such name embeddings, all variants (e.g., from Civitai) derived from the same base model (i.e., SDXL) readily become identity-aware text-to-image models. Project homepage: \url{https://magicfusion.github.io/MagicNaming/}.

Xiuyuan Chen, Tao Sun, Dexin Su et al.

Current benchmarks for AI clinician systems, often based on multiple-choice exams or manual rubrics, fail to capture the depth, robustness, and safety required for real-world clinical practice. To address this, we introduce the GAPS framework, a multidimensional paradigm for evaluating \textbf{G}rounding (cognitive depth), \textbf{A}dequacy (answer completeness), \textbf{P}erturbation (robustness), and \textbf{S}afety. Critically, we developed a fully automated, guideline-anchored pipeline to construct a GAPS-aligned benchmark end-to-end, overcoming the scalability and subjectivity limitations of prior work. Our pipeline assembles an evidence neighborhood, creates dual graph and tree representations, and automatically generates questions across G-levels. Rubrics are synthesized by a DeepResearch agent that mimics GRADE-consistent, PICO-driven evidence review in a ReAct loop. Scoring is performed by an ensemble of large language model (LLM) judges. Validation confirmed our automated questions are high-quality and align with clinician judgment. Evaluating state-of-the-art models on the benchmark revealed key failure modes: performance degrades sharply with increased reasoning depth (G-axis), models struggle with answer completeness (A-axis), and they are highly vulnerable to adversarial perturbations (P-axis) as well as certain safety issues (S-axis). This automated, clinically-grounded approach provides a reproducible and scalable method for rigorously evaluating AI clinician systems and guiding their development toward safer, more reliable clinical practice.

Xiefeng Wu, Jing Zhao, Shu Zhang et al.

Online reinforcement learning in complex tasks is time-consuming, as massive interaction steps are needed to learn the optimal Q-function.Vision-language action (VLA) policies represent a promising direction for solving diverse tasks; however, their performance on low-level control remains limited, and effective deployment often requires task-specific expert demonstrations for fine-tuning. In this paper, we propose \textbf{VARL} (\textbf{V}LM as \textbf{A}ction advisor for online \textbf{R}einforcement \textbf{L}earning), a framework that leverages the domain knowledge of vision-language models (VLMs) to provide action suggestions for reinforcement learning agents. Unlike previous methods, VARL provides action suggestions rather than designing heuristic rewards, thereby guaranteeing unchanged optimality and convergence. The suggested actions increase sample diversity and ultimately improve sample efficiency, especially in sparse-reward tasks. To validate the effectiveness of VARL, we evaluate it across diverse environments and agent settings. Results show that VARL greatly improves sample efficiency without introducing significant computational overhead. These advantages make VARL a general framework for online reinforcement learning and make it feasible to directly apply reinforcement learning from scratch in real-world environments.

Chenyu Lu, Shiliang Sun, Jing Zhao et al.

Multimodal machine translation (MMT) seeks to address the challenges posed by linguistic polysemy and ambiguity in translation tasks by incorporating visual information. A key bottleneck in current MMT research is the effective utilization of visual data. Previous approaches have focused on extracting global or region-level image features and using attention or gating mechanisms for multimodal information fusion. However, these methods have not adequately tackled the issue of visual information redundancy in MMT, nor have they proposed effective solutions. In this paper, we introduce a novel approach--multimodal machine translation with visual Scene Graph Pruning (PSG), which leverages language scene graph information to guide the pruning of redundant nodes in visual scene graphs, thereby reducing noise in downstream translation tasks. Through extensive comparative experiments with state-of-the-art methods and ablation studies, we demonstrate the effectiveness of the PSG model. Our results also highlight the promising potential of visual information pruning in advancing the field of MMT.

Zhuang Yu, Shiliang Sun, Jing Zhao et al.

Multimodal Machine Translation (MMT) aims to improve translation quality by leveraging auxiliary modalities such as images alongside textual input. While recent advances in large-scale pre-trained language and vision models have significantly benefited unimodal natural language processing tasks, their effectiveness and role in MMT remain underexplored. In this work, we conduct a systematic study on the impact of pre-trained encoders and decoders in multimodal translation models. Specifically, we analyze how different training strategies, from training from scratch to using pre-trained and partially frozen components, affect translation performance under a unified MMT framework. Experiments are carried out on the Multi30K and CoMMuTE dataset across English-German and English-French translation tasks. Our results reveal that pre-training plays a crucial yet asymmetrical role in multimodal settings: pre-trained decoders consistently yield more fluent and accurate outputs, while pre-trained encoders show varied effects depending on the quality of visual-text alignment. Furthermore, we provide insights into the interplay between modality fusion and pre-trained components, offering guidance for future architecture design in multimodal translation systems.

Ran Chen, Zeke Lian, Yueheng He et al.

In traditional human living environment landscape design, the establishment of three-dimensional models is an essential step for designers to intuitively present the spatial relationships of design elements, as well as a foundation for conducting landscape analysis on the site. Rapidly and effectively generating beautiful and realistic landscape spaces is a significant challenge faced by designers. Although generative design has been widely applied in related fields, they mostly generate three-dimensional models through the restriction of indicator parameters. However, the elements of landscape design are complex and have unique requirements, making it difficult to generate designs from the perspective of indicator limitations. To address these issues, this study proposes a park space generative design system based on deep learning technology. This system generates design plans based on the topological relationships of landscape elements, then vectorizes the plan element information, and uses Grasshopper to generate three-dimensional models while synchronously fine-tuning parameters, rapidly completing the entire process from basic site conditions to model effect analysis. Experimental results show that: (1) the system, with the aid of AI-assisted technology, can rapidly generate space green space schemes that meet the designer's perspective based on site conditions; (2) this study has vectorized and three-dimensionalized various types of landscape design elements based on semantic information; (3) the analysis and visualization module constructed in this study can perform landscape analysis on the generated three-dimensional models and produce node effect diagrams, allowing users to modify the design in real time based on the effects, thus enhancing the system's interactivity.

Jing Zhao

Due to the influence of imaging equipment and complex imaging environments, most images in daily life have features of intensity inhomogeneity and noise. Therefore, many scholars have designed many image segmentation algorithms to address these issues. Among them, the active contour model is one of the most effective image segmentation algorithms.This paper proposes an active contour model driven by the hybrid signed pressure function that combines global and local information construction. Firstly, a new global region-based signed pressure function is introduced by combining the average intensity of the inner and outer regions of the curve with the median intensity of the inner region of the evolution curve. Then, the paper uses the energy differences between the inner and outer regions of the curve in the local region to design the signed pressure function of the local term. Combine the two SPF function to obtain a new signed pressure function and get the evolution equation of the new model. Finally, experiments and numerical analysis show that the model has excellent segmentation performance for both intensity inhomogeneous images and noisy images.

Ran Chen, Xueqi Yao, Jing Zhao et al.

This study aims to explore the complex relationship between perceptual and cognitive interactions in multimodal data analysis,with a specific emphasis on spatial experience design in overseas Chinese gardens. It is found that evaluation content and images on social media can reflect individuals' concerns and sentiment responses, providing a rich data base for cognitive research that contains both sentimental and image-based cognitive information. Leveraging deep learning techniques, we analyze textual and visual data from social media, thereby unveiling the relationship between people's perceptions and sentiment cognition within the context of overseas Chinese gardens. In addition, our study introduces a multi-agent system (MAS)alongside AI agents. Each agent explores the laws of aesthetic cognition through chat scene simulation combined with web search. This study goes beyond the traditional approach of translating perceptions into sentiment scores, allowing for an extension of the research methodology in terms of directly analyzing texts and digging deeper into opinion data. This study provides new perspectives for understanding aesthetic experience and its impact on architecture and landscape design across diverse cultural contexts, which is an essential contribution to the field of cultural communication and aesthetic understanding.

Gemini Team, Rohan Anil, Sebastian Borgeaud et al.

This report introduces a new family of multimodal models, Gemini, that exhibit remarkable capabilities across image, audio, video, and text understanding. The Gemini family consists of Ultra, Pro, and Nano sizes, suitable for applications ranging from complex reasoning tasks to on-device memory-constrained use-cases. Evaluation on a broad range of benchmarks shows that our most-capable Gemini Ultra model advances the state of the art in 30 of 32 of these benchmarks - notably being the first model to achieve human-expert performance on the well-studied exam benchmark MMLU, and improving the state of the art in every one of the 20 multimodal benchmarks we examined. We believe that the new capabilities of the Gemini family in cross-modal reasoning and language understanding will enable a wide variety of use cases. We discuss our approach toward post-training and deploying Gemini models responsibly to users through services including Gemini, Gemini Advanced, Google AI Studio, and Cloud Vertex AI.

Ran Chen, Xingjian Yi, Jing Zhao et al.

The development of generative design driven by artificial intelligence algorithms is speedy. There are two research gaps in the current research: 1) Most studies only focus on the relationship between design elements and pay little attention to the external information of the site; 2) GAN and other traditional generative algorithms generate results with low resolution and insufficient details. To address these two problems, we integrate GAN, Stable diffusion multimodal large-scale image pre-training model to construct a full-process park generative design method: 1) First, construct a high-precision remote sensing object extraction system for automated extraction of urban environmental information; 2) Secondly, use GAN to construct a park design generation system based on the external environment, which can quickly infer and generate design schemes from urban environmental information; 3) Finally, introduce Stable Diffusion to optimize the design plan, fill in details, and expand the resolution of the plan by 64 times. This method can achieve a fully unmanned design automation workflow. The research results show that: 1) The relationship between the inside and outside of the site will affect the algorithm generation results. 2) Compared with traditional GAN algorithms, Stable diffusion significantly improve the information richness of the generated results.

Zihao Wang, Jing Zhao, Xuetong Ding et al.

Face plays an important role in humans visual perception, and reconstructing perceived faces from brain activities is challenging because of its difficulty in extracting high-level features and maintaining consistency of multiple face attributes, such as expression, identity, gender, etc. In this study, we proposed a novel reconstruction framework, which we called Double-Flow GAN, that can enhance the capability of discriminator and handle imbalances in images from certain domains that are too easy for generators. We also designed a pretraining process that uses features extracted from images as conditions for making it possible to pretrain the conditional reconstruction model from fMRI in a larger pure image dataset. Moreover, we developed a simple pretrained model for fMRI alignment to alleviate the problem of cross-subject reconstruction due to the variations of brain structure among different subjects. We conducted experiments by using our proposed method and traditional reconstruction models. Results showed that the proposed method is significant at accurately reconstructing multiple face attributes, outperforms the previous reconstruction models, and exhibited state-of-the-art reconstruction abilities.

Jing Zhao, Heliang Zheng, Chaoyue Wang et al.

Classifier-free guidance is an effective sampling technique in diffusion models that has been widely adopted. The main idea is to extrapolate the model in the direction of text guidance and away from null-text guidance. In this paper, we demonstrate that null-text guidance in diffusion models is secretly a cartoon-style creator, i.e., the generated images can be efficiently transformed into cartoons by simply perturbing the null-text guidance. Specifically, we proposed two disturbance methods, i.e., Rollback disturbance (Back-D) and Image disturbance (Image-D), to construct misalignment between the noisy images used for predicting null-text guidance and text guidance (subsequently referred to as \textbf{null-text noisy image} and \textbf{text noisy image} respectively) in the sampling process. Back-D achieves cartoonization by altering the noise level of null-text noisy image via replacing $x_t$ with $x_{t+Δt}$. Image-D, alternatively, produces high-fidelity, diverse cartoons by defining $x_t$ as a clean input image, which further improves the incorporation of finer image details. Through comprehensive experiments, we delved into the principle of noise disturbing for null-text and uncovered that the efficacy of disturbance depends on the correlation between the null-text noisy image and the source image. Moreover, our proposed techniques, which can generate cartoon images and cartoonize specific ones, are training-free and easily integrated as a plug-and-play component in any classifier-free guided diffusion model. Project page is available at \url{https://nulltextforcartoon.github.io/}.

Jing Zhao, Junwei Bao, Yifan Wang et al.

Transformer-based pre-trained models, such as BERT, have achieved remarkable results on machine reading comprehension. However, due to the constraint of encoding length (e.g., 512 WordPiece tokens), a long document is usually split into multiple chunks that are independently read. It results in the reading field being limited to individual chunks without information collaboration for long document machine reading comprehension. To address this problem, we propose RoR, a read-over-read method, which expands the reading field from chunk to document. Specifically, RoR includes a chunk reader and a document reader. The former first predicts a set of regional answers for each chunk, which are then compacted into a highly-condensed version of the original document, guaranteeing to be encoded once. The latter further predicts the global answers from this condensed document. Eventually, a voting strategy is utilized to aggregate and rerank the regional and global answers for final prediction. Extensive experiments on two benchmarks QuAC and TriviaQA demonstrate the effectiveness of RoR for long document reading. Notably, RoR ranks 1st place on the QuAC leaderboard (https://quac.ai/) at the time of submission (May 17th, 2021).

Jing Zhao, Jingya Wang, Madhav Sigdel et al.

The online recruitment matching system has been the core technology and service platform in CareerBuilder. One of the major challenges in an online recruitment scenario is to provide good matches between job posts and candidates using a recommender system on the scale. In this paper, we discussed the techniques for applying an embedding-based recommender system for the large scale of job to candidates matching. To learn the comprehensive and effective embedding for job posts and candidates, we have constructed a fused-embedding via different levels of representation learning from raw text, semantic entities and location information. The clusters of fused-embedding of job and candidates are then used to build and train the Faiss index that supports runtime approximate nearest neighbor search for candidate retrieval. After the first stage of candidate retrieval, a second stage reranking model that utilizes other contextual information was used to generate the final matching result. Both offline and online evaluation results indicate a significant improvement of our proposed two-staged embedding-based system in terms of click-through rate (CTR), quality and normalized discounted accumulated gain (nDCG), compared to those obtained from our baseline system. We further described the deployment of the system that supports the million-scale job and candidate matching process at CareerBuilder. The overall improvement of our job to candidate matching system has demonstrated its feasibility and scalability at a major online recruitment site.

Yachen Tang, Haiyun Han, Xianmao Yu et al.

The intelligent question answering (IQA) system can accurately capture users' search intention by understanding the natural language questions, searching relevant content efficiently from a massive knowledge-base, and returning the answer directly to the user. Since the IQA system can save inestimable time and workforce in data search and reasoning, it has received more and more attention in data science and artificial intelligence. This article introduced a domain knowledge graph using the graph database and graph computing technologies from massive heterogeneous data in electric power. It then proposed an IQA system based on the electrical power knowledge graph to extract the intent and constraints of natural interrogation based on the natural language processing (NLP) method, to construct graph data query statements via knowledge reasoning, and to complete the accurate knowledge search and analysis to provide users with an intuitive visualization. This method thoroughly combined knowledge graph and graph computing characteristics, realized high-speed multi-hop knowledge correlation reasoning analysis in tremendous knowledge. The proposed work can also provide a basis for the context-aware intelligent question and answer.

Jing Zhao, Junwei Bao, Yifan Wang et al.

Keyphrases, that concisely summarize the high-level topics discussed in a document, can be categorized into present keyphrase which explicitly appears in the source text, and absent keyphrase which does not match any contiguous subsequence but is highly semantically related to the source. Most existing keyphrase generation approaches synchronously generate present and absent keyphrases without explicitly distinguishing these two categories. In this paper, a Select-Guide-Generate (SGG) approach is proposed to deal with present and absent keyphrase generation separately with different mechanisms. Specifically, SGG is a hierarchical neural network which consists of a pointing-based selector at low layer concentrated on present keyphrase generation, a selection-guided generator at high layer dedicated to absent keyphrase generation, and a guider in the middle to transfer information from selector to generator. Experimental results on four keyphrase generation benchmarks demonstrate the effectiveness of our model, which significantly outperforms the strong baselines for both present and absent keyphrases generation. Furthermore, we extend SGG to a title generation task which indicates its extensibility in natural language generation tasks.

Jing Zhao, Samanvitha Basole, Mark Stamp

Discrete hidden Markov models (HMM) are often applied to malware detection and classification problems. However, the continuous analog of discrete HMMs, that is, Gaussian mixture model-HMMs (GMM-HMM), are rarely considered in the field of cybersecurity. In this paper, we use GMM-HMMs for malware classification and we compare our results to those obtained using discrete HMMs. As features, we consider opcode sequences and entropy-based sequences. For our opcode features, GMM-HMMs produce results that are comparable to those obtained using discrete HMMs, whereas for our entropy-based features, GMM-HMMs generally improve significantly on the classification results that we have achieved with discrete HMMs.

Zhiqian Qiao, Jing Zhao, Zachariah Tyree et al.

How autonomous vehicles and human drivers share public transportation systems is an important problem, as fully automatic transportation environments are still a long way off. Understanding human drivers' behavior can be beneficial for autonomous vehicle decision making and planning, especially when the autonomous vehicle is surrounded by human drivers who have various driving behaviors and patterns of interaction with other vehicles. In this paper, we propose an LSTM-based trajectory prediction method for human drivers which can help the autonomous vehicle make better decisions, especially in urban intersection scenarios. Meanwhile, in order to collect human drivers' driving behavior data in the urban scenario, we describe a system called UrbanFlow which includes the whole procedure from raw bird's-eye view data collection via drone to the final processed trajectories. The system is mainly intended for urban scenarios but can be extended to be used for any traffic scenarios.

Athar Sefid, Jian Wu, Allen C. Ge et al.

Automatically extracted metadata from scholarly documents in PDF formats is usually noisy and heterogeneous, often containing incomplete fields and erroneous values. One common way of cleaning metadata is to use a bibliographic reference dataset. The challenge is to match records between corpora with high precision. The existing solution which is based on information retrieval and string similarity on titles works well only if the titles are cleaned. We introduce a system designed to match scholarly document entities with noisy metadata against a reference dataset. The blocking function uses the classic BM25 algorithm to find the matching candidates from the reference data that has been indexed by ElasticSearch. The core components use supervised methods which combine features extracted from all available metadata fields. The system also leverages available citation information to match entities. The combination of metadata and citation achieves high accuracy that significantly outperforms the baseline method on the same test dataset. We apply this system to match the database of CiteSeerX against Web of Science, PubMed, and DBLP. This method will be deployed in the CiteSeerX system to clean metadata and link records to other scholarly big datasets.

Shiliang Sun, Zehui Cao, Han Zhu et al.

Machine learning develops rapidly, which has made many theoretical breakthroughs and is widely applied in various fields. Optimization, as an important part of machine learning, has attracted much attention of researchers. With the exponential growth of data amount and the increase of model complexity, optimization methods in machine learning face more and more challenges. A lot of work on solving optimization problems or improving optimization methods in machine learning has been proposed successively. The systematic retrospect and summary of the optimization methods from the perspective of machine learning are of great significance, which can offer guidance for both developments of optimization and machine learning research. In this paper, we first describe the optimization problems in machine learning. Then, we introduce the principles and progresses of commonly used optimization methods. Next, we summarize the applications and developments of optimization methods in some popular machine learning fields. Finally, we explore and give some challenges and open problems for the optimization in machine learning.

Jing Zhao, Jingjing Fei, Shiliang Sun

In order to better model high-dimensional sequential data, we propose a collaborative multi-output Gaussian process dynamical system (CGPDS), which is a novel variant of GPDSs. The proposed model assumes that the output on each dimension is controlled by a shared global latent process and a private local latent process. Thus, the dependence among different dimensions of the sequences can be captured, and the unique characteristics of each dimension of the sequences can be maintained. For training models and making prediction, we introduce inducing points and adopt stochastic variational inference methods.

Jing Zhao, Lili Miao, Haiyang Fang et al.

Link prediction aims to uncover missing links or predict the emergence of future relationships according to the current networks structure. Plenty of algorithms have been developed for link prediction in unweighted networks, with only a very few of them having been extended to weighted networks. Thus far, how to predict weights of links is important but rarely studied. In this Letter, we present a reliable-route-based method to extend unweighted local similarity indices to weighted indices and propose a method to predict both the link existence and link weights accordingly. Experiments on different real networks suggest that the weighted resource allocation index has the best performance to predict the existence of links, while the reliable-route-based weighted resource allocation index performs noticeably better on weight prediction. Further analysis shows a strong correlation for both link prediction and weight prediction: the larger the clustering coefficient, the higher the prediction accuracy.