2 citations · 5 across the 24 of their papers we have counts for
24 papers
1% of Tokens Can Be Enough: On Gradient Estimation in On-Policy Distillation
Huanxin Sheng, Zhiling Ye, Haonan Wang +3
Sparse on-policy distillation (OPD) allocates teacher supervision to a small subset of tokens in student-generated trajectories. However, useful teacher guidance can yield a noisy…
SimpleSearch-VL: A Simple Recipe for Multimodal Agentic Deep Search
Ming Dai, Zhihong Lu, Jinjie Gu +5
We present SimpleSearch-VL, an efficient, reliable, and practical framework for multimodal agentic search. Its core idea is to improve the agent's own search-and-verification proce…
LiveAgentBench: Comprehensive Benchmarking of Agentic Systems Across 104 Real-World Challenges
Hao Li, Huan Wang, Jinjie Gu +3
As large language models grow more capable, general AI agents have become increasingly prevalent in practical applications. However, existing benchmarks face significant limitation…
V2P: Visual Attention Calibration for GUI Grounding via Background Suppression and Center Peaking
Jikai Chen, Long Chen, Dong Wang +6
Precise localization of GUI elements is crucial for the development of GUI agents. Traditional methods rely on bounding box or center-point regression, neglecting spatial interacti…
FunReason-MT Technical Report: Advanced Data Synthesis Solution for Real-world Multi-Turn Tool-use
Zengzhuang Xu, Bingguang Hao, Zechuan Wang +14
Function calling (FC) empowers large language models (LLMs) and autonomous agents to interface with external tools, a critical capability for solving complex, real-world problems.…
Don't Just Fine-tune the Agent, Tune the Environment
Siyuan Lu, Zechuan Wang, Hongxuan Zhang +5
Large Language Model (LLM) agents show great promise for complex, multi-turn tool-use tasks, but their development is often hampered by the extreme scarcity of high-quality trainin…
Recon-Act: A Self-Evolving Multi-Agent Browser-Use System via Web Reconnaissance, Tool Generation, and Task Execution
Kaiwen He, Zhiwei Wang, Chenyi Zhuang +1
Recent years, multimodal models have made remarkable strides and pave the way for intelligent browser use agents. However, when solving tasks on real world webpages in multi-turn,…
AWorld: Orchestrating the Training Recipe for Agentic AI
Chengyue Yu, Siyuan Lu, Chenyi Zhuang +14
The learning from practice paradigm is crucial for developing capable Agentic AI systems, yet it is severely hampered by inefficient experience generation, a bottleneck especially…
Profile-Aware Maneuvering: A Dynamic Multi-Agent System for Robust GAIA Problem Solving by AWorld
Zhitian Xie, Qintong Wu, Chengyue Yu +2
The rapid advancement of large language models (LLMs) has empowered intelligent agents to leverage diverse external tools for solving complex real-world problems. However, this rel…
V2P: Visual Attention Calibration for GUI Grounding via Background Suppression and Center Peaking
Jikai Chen, Long Chen, Dong Wang +6
Precise localization of GUI elements is crucial for the development of GUI agents. Traditional methods rely on bounding box or center-point regression, neglecting spatial interacti…
Reasoning through Exploration: A Reinforcement Learning Framework for Robust Function Calling
Bingguang Hao, Zengzhuang Xu, Maolin Wang +9
The effective training of Large Language Models (LLMs) for function calling faces a critical challenge: balancing exploration of complex reasoning paths with stable policy optimiza…
RAG-R1: Incentivizing the Search and Reasoning Capabilities of LLMs through Multi-query Parallelism
Zhiwen Tan, Jiaming Huang, Qintong Wu +3
Large Language Models (LLMs), despite their remarkable capabilities, are prone to generating hallucinated or outdated content due to their static internal knowledge. While Retrieva…
FunReason: Enhancing Large Language Models' Function Calling via Self-Refinement Multiscale Loss and Automated Data Refinement
Bingguang Hao, ZengZhuang Xu, Maolin Wang +9
The integration of large language models (LLMs) with function calling has emerged as a crucial capability for enhancing their practical utility in real-world applications. However,…
A Comprehensive Survey of Reward Models: Taxonomy, Applications, Challenges, and Future
Jialun Zhong, Wei Shen, Yanzeng Li +7
Reward Model (RM) has demonstrated impressive potential for enhancing Large Language Models (LLM), as RM can serve as a proxy for human preferences, providing signals to guide LLMs…
CSR:Achieving 1 Bit Key-Value Cache via Sparse Representation
Hongxuan Zhang, Yao Zhao, Jiaqi Zheng +3
The emergence of long-context text applications utilizing large language models (LLMs) has presented significant scalability challenges, particularly in memory footprint. The linea…
Explainable Behavior Cloning: Teaching Large Language Model Agents through Learning by Demonstration
Yanchu Guan, Dong Wang, Yan Wang +5
Autonomous mobile app interaction has become increasingly important with growing complexity of mobile applications. Developing intelligent agents that can effectively navigate and…
Mitigate Position Bias with Coupled Ranking Bias on CTR Prediction
Yao Zhao, Zhining Liu, Tianchi Cai +3
Position bias, i.e., users' preference of an item is affected by its placing position, is well studied in the recommender system literature. However, most existing methods ignore t…
DNS-Rec: Data-aware Neural Architecture Search for Recommender Systems
Sheng Zhang, Maolin Wang, Yao Zhao +6
In the era of data proliferation, efficiently sifting through vast information to extract meaningful insights has become increasingly crucial. This paper addresses the computationa…
MoDE: A Mixture-of-Experts Model with Mutual Distillation among the Experts
Zhitian Xie, Yinger Zhang, Chenyi Zhuang +4
The application of mixture-of-experts (MoE) is gaining popularity due to its ability to improve model's performance. In an MoE structure, the gate layer plays a significant role in…
CharPoet: A Chinese Classical Poetry Generation System Based on Token-free LLM
Chengyue Yu, Lei Zang, Jiaotuan Wang +2
Automatic Chinese classical poetry generation has attracted much research interest, but achieving effective control over format and content simultaneously remains challenging. Trad…
Intelligent Virtual Assistants with LLM-based Process Automation
Yanchu Guan, Dong Wang, Zhixuan Chu +6
While intelligent virtual assistants like Siri, Alexa, and Google Assistant have become ubiquitous in modern life, they still face limitations in their ability to follow multi-step…
Lookahead: An Inference Acceleration Framework for Large Language Model with Lossless Generation Accuracy
Yao Zhao, Zhitian Xie, Chen Liang +2
As Large Language Models (LLMs) have made significant advancements across various tasks, such as question answering, translation, text summarization, and dialogue systems, the need…
Large Multimodal Model Compression via Efficient Pruning and Distillation at AntGroup
Maolin Wang, Yao Zhao, Jiajia Liu +5
The deployment of Large Multimodal Models (LMMs) within AntGroup has significantly advanced multimodal tasks in payment, security, and advertising, notably enhancing advertisement…
Fast Chain-of-Thought: A Glance of Future from Parallel Decoding Leads to Answers Faster
Hongxuan Zhang, Zhining Liu, Yao Zhao +4
In this work, we propose FastCoT, a model-agnostic framework based on parallel decoding without any further training of an auxiliary model or modification to the LLM itself. FastCo…