21 papers
Enhancing Social Intelligence in LLMs with Hierarchical Reasoning and Utterance-Level Goal Rewarding
Xiaofeng Wang, Kakam Chong, Shuai Xiao +9
Large language models (LLMs) excel in structured tasks but struggle with dynamic social interactions, where success requires long-term goal coordination and rapid adaptation. Curre…
MemoryDocDataSet: A Benchmark for Joint Conversational Memory and Long Document Reasoning
Qiyang Xie, Jialun Wu, Xinjie He +4
AI systems increasingly need to combine two demanding capabilities: navigating multi-session conversation history and performing deep reading comprehension within long documents. Y…
When Does Persona Prompting Actually Help? A Retrieval and Metric Analysis of Expert Role Injection in LLMs
Shuai Xiao, Su Liu, Weikai Zhou +4
Persona prompting is widely used to steer large language models, yet its practical value remains unclear. Prior work often evaluates persona prompting using aggregate scores, makin…
What Training Data Teaches RL Memory Agents: An Empirical Study of Curriculum Effects in Memory-Augmented QA
Xinjie He, Zhiyuan Lin, Su Liu +4
Reinforcement learning (RL) has emerged as a viable recipe for training LLM agents to reason over external memory banks in multi-session dialogue. Existing work trains exclusively…
Geometrically Constrained Stenosis Editing in Coronary Angiography via Entropic Optimal Transport
Jialin Li, Zhuo Zhang, Yue Cao +4
The scarcity of high-quality imaging data for coronary angiography (CAG) stenosis limits the clinical translation of automated stenosis detection. Synthetic stenosis data provides…
Improving Human Image Animation via Semantic Representation Alignment
Chang Liu, Mengting Chen, Yixuan Huang +5
The field of image-to-video generation has made remarkable progress. However, challenges such as human limb twisting and facial distortion persist, especially when generating long…
Explore More, Learn Better: Parallel MLLM Embeddings under Mutual Information Minimization
Zhicheng Wang, Chen Ju, Xu Chen +5
Embedding models are a cornerstone of modern AI. Driven by Multimodal Large Language Models (MLLMs), they have made great progress in architecture and data curation, while the holi…
Wave-Particle (Continuous-Discrete) Dualistic Visual Tokenization for Unified Understanding and Generation
Yizhu Chen, Chen Ju, Zhicheng Wang +5
The unification of understanding and generation within a single multi-modal large model (MLLM) remains one significant challenge, largely due to the dichotomy between continuous an…
Leveraging Scene Context with Dual Networks for Sequential User Behavior Modeling
Xu Chen, Yunmeng Shu, Yuangang Pan +6
Modeling sequential user behaviors for future behavior prediction is crucial in improving user's information retrieval experience. Recent studies highlight the importance of incorp…
Evaluating Multi-Turn Bargain Skills in LLM-Based Seller Agent
Issue Yishu Wang, Kakam Chong, Xiaofeng Wang +7
In online second-hand marketplaces, multi-turn bargaining is a crucial part of seller-buyer interactions. Large Language Models (LLMs) can act as seller agents, negotiating with bu…
MMKB-RAG: A Multi-Modal Knowledge-Based Retrieval-Augmented Generation Framework
Zihan Ling, Zhiyao Guo, Yixuan Huang +5
Recent advancements in large language models (LLMs) and multi-modal LLMs have been remarkable. However, these models still rely solely on their parametric knowledge, which limits t…
Beyond Static Scenes: Camera-controllable Background Generation for Human Motion
Mingshuai Yao, Mengting Chen, Qinye Zhou +9
In this paper, we investigate the generation of new video backgrounds given a human foreground video, a camera pose, and a reference scene image. This task presents three key chall…
Squeeze Out Tokens from Sample for Finer-Grained Data Governance
Weixiong Lin, Chen Ju, Haicheng Wang +8
Widely observed data scaling laws, in which error falls off as a power of the training size, demonstrate the diminishing returns of unselective data expansion. Hence, data governan…
Contrast-Unity for Partially-Supervised Temporal Sentence Grounding
Haicheng Wang, Chen Ju, Weixiong Lin +4
Temporal sentence grounding aims to detect event timestamps described by the natural language query from given untrimmed videos. The existing fully-supervised setting achieves grea…
Advancing Myopia To Holism: Fully Contrastive Language-Image Pre-training
Haicheng Wang, Chen Ju, Weixiong Lin +9
In rapidly evolving field of vision-language models (VLMs), contrastive language-image pre-training (CLIP) has made significant strides, becoming foundation for various downstream…
Counterfactual Learning-Driven Representation Disentanglement for Search-Enhanced Recommendation
Jiajun Cui, Xu Chen, Shuai Xiao +4
For recommender systems in internet platforms, search activities provide additional insights into user interest through query-click interactions with items, and are thus widely use…
Learning Multi-Branch Cooperation for Enhanced Click-Through Rate Prediction at Taobao
Xu Chen, Zida Cheng, Yuangang Pan +6
Existing click-through rate (CTR) prediction works have studied the role of feature interaction through a variety of techniques. Each interaction technique exhibits its own strengt…
Turbo: Informativity-Driven Acceleration Plug-In for Vision-Language Large Models
Chen Ju, Haicheng Wang, Haozhe Cheng +6
Vision-Language Large Models (VLMs) recently become primary backbone of AI, due to the impressive performance. However, their expensive computation costs, i.e., throughput and dela…
Tunnel Try-on: Excavating Spatial-temporal Tunnels for High-quality Virtual Try-on in Videos
Zhengze Xu, Mengting Chen, Zhao Wang +6
Video try-on is a challenging task and has not been well tackled in previous works. The main obstacle lies in preserving the details of the clothing and modeling the coherent motio…
Cell Variational Information Bottleneck Network
Zhonghua Zhai, Chen Ju, Jinsong Lan +1
In this work, we propose Cell Variational Information Bottleneck Network (cellVIB), a convolutional neural network using information bottleneck mechanism, which can be combined wit…
Wear-Any-Way: Manipulable Virtual Try-on via Sparse Correspondence Alignment
Mengting Chen, Xi Chen, Zhonghua Zhai +4
This paper introduces a novel framework for virtual try-on, termed Wear-Any-Way. Different from previous methods, Wear-Any-Way is a customizable solution. Besides generating high-f…