12 citations · 36 across the 21 of their papers we have counts for
22 papers
SOAR: Source-Aware Reinforcement Learning with Precise Credit Assignment for Search Agents
Shengze Huang, Zihao Tang, Jinluan Yang +11
Reinforcement learning (RL) for search agents often suffers from evidence fabrication, a failure mode where agents bypass external information () and fabricate an incorrec…
Group-Graph Policy Optimization for Long-Horizon Agentic Reinforcement Learning
Yunan Wang, Minghui Song, Zihan Zhang +6
Group-based Reinforcement Learning (RL) has significantly enhanced Large Language Models (LLMs) in agentic scenarios. To achieve finer-grained policy updates, recent agentic RL fra…
Beyond Static Dialogues: Benchmarking Realistic, Heterogeneous, and Evolving Long-Term Memory
Han Zhang, Zihao Tang, Xin Yu +8
In existing memory benchmarks for Large Language Models (LLMs), the evaluated dialogue sessions often lack long-term semantic consistency, and the underlying personas tend to be fl…
Mnemis: Dual-Route Retrieval on Hierarchical Graphs for Long-Term LLM Memory
Zihao Tang, Xin Yu, Ziyu Xiao +9
AI Memory, specifically how models organizes and retrieves historical messages, becomes increasingly valuable to Large Language Models (LLMs), yet existing methods (RAG and Graph-R…
TL;DR: Too Long, Do Re-weighting for Efficient LLM Reasoning Compression
Zhong-Zhi Li, Xiao Liang, Zihao Tang +11
Large Language Models (LLMs) have recently achieved remarkable progress by leveraging Reinforcement Learning and extended Chain-of-Thought (CoT) techniques. However, the challenge…
Context-DPO: Aligning Language Models for Context-Faithfulness
Baolong Bi, Shaohan Huang, Yiwei Wang +11
Reliable responses from large language models (LLMs) require adherence to user instructions and retrieved information. While alignment techniques help LLMs align with human intenti…
E5-V: Universal Embeddings with Multimodal Large Language Models
Ting Jiang, Minghui Song, Zihan Zhang +6
Multimodal large language models (MLLMs) have shown promising advancements in general visual and language understanding. However, the representation of multimodal information using…
ASI++: Towards Distributionally Balanced End-to-End Generative Retrieval
Yuxuan Liu, Tianchi Yang, Zihan Zhang +5
Generative retrieval, a promising new paradigm in information retrieval, employs a seq2seq model to encode document features into parameters and decode relevant document identifier…
MoRA: High-Rank Updating for Parameter-Efficient Fine-Tuning
Ting Jiang, Shaohan Huang, Shengyue Luo +8
Low-rank adaptation is a popular parameter-efficient fine-tuning method for large language models. In this paper, we analyze the impact of low-rank updating, as implemented in LoRA…
ResLoRA: Identity Residual Mapping in Low-Rank Adaption
Shuhua Shi, Shaohan Huang, Minghui Song +7
As one of the most popular parameter-efficient fine-tuning (PEFT) methods, low-rank adaptation (LoRA) is commonly applied to fine-tune large language models (LLMs). However, updati…
HD-Eval: Aligning Large Language Model Evaluators Through Hierarchical Criteria Decomposition
Yuxuan Liu, Tianchi Yang, Shaohan Huang +6
Large language models (LLMs) have emerged as a promising alternative to expensive human evaluations. However, the alignment and coverage of LLM-based evaluations are often limited…
Text Diffusion with Reinforced Conditioning
Yuxuan Liu, Tianchi Yang, Shaohan Huang +6
Diffusion models have demonstrated exceptional capability in generating high-quality images, videos, and audio. Due to their adaptiveness in iterative refinement, they provide a st…
Improving Domain Adaptation through Extended-Text Reading Comprehension
Ting Jiang, Shaohan Huang, Shengyue Luo +8
To enhance the domain-specific capabilities of large language models, continued pre-training on a domain-specific corpus is a prevalent method. Recent work demonstrates that adapti…
Auto Search Indexer for End-to-End Document Retrieval
Tianchi Yang, Minghui Song, Zihan Zhang +4
Generative retrieval, which is a new advanced paradigm for document retrieval, has recently attracted research interests, since it encodes all documents into the model and directly…
Democratizing Reasoning Ability: Tailored Learning from Large Language Model
Zhaoyang Wang, Shaohan Huang, Yuxuan Liu +8
Large language models (LLMs) exhibit impressive emergent abilities in natural language processing, but their democratization is hindered due to huge computation requirements and cl…
Calibrating LLM-Based Evaluator
Yuxuan Liu, Tianchi Yang, Shaohan Huang +6
Recent advancements in large language models (LLMs) on language modeling and emergent capabilities make them a promising reference-free evaluator of natural language generation qua…
Towards Better Entity Linking with Multi-View Enhanced Distillation
Yi Liu, Yuan Tian, Jianxun Lian +7
Dense retrieval is widely used for entity linking to retrieve entities from large-scale knowledge bases. Mainstream techniques are based on a dual-encoder framework, which encodes…
Dual-Alignment Pre-training for Cross-lingual Sentence Embedding
Ziheng Li, Shaohan Huang, Zihan Zhang +7
Recent studies have shown that dual encoder models trained with the sentence-level translation ranking task are effective methods for cross-lingual sentence embedding. However, our…
Pre-training Language Model as a Multi-perspective Course Learner
Beiduo Chen, Shaohan Huang, Zihan Zhang +6
ELECTRA, the generator-discriminator pre-training framework, has achieved impressive semantic construction capability among various downstream tasks. Despite the convincing perform…
RAPO: An Adaptive Ranking Paradigm for Bilingual Lexicon Induction
Zhoujin Tian, Chaozhuo Li, Shuo Ren +8
Bilingual lexicon induction induces the word translations by aligning independently trained word embeddings in two languages. Existing approaches generally focus on minimizing the…
PromptBERT: Improving BERT Sentence Embeddings with Prompts
Ting Jiang, Jian Jiao, Shaohan Huang +7
We propose PromptBERT, a novel contrastive learning method for learning better sentence representation. We firstly analyze the drawback of current sentence embedding from original…
Improving Non-autoregressive Generation with Mixup Training
Ting Jiang, Shaohan Huang, Zihan Zhang +6
While pre-trained language models have achieved great success on various natural language understanding tasks, how to effectively leverage them into non-autoregressive generation t…