Papers by Kai Lv
FastMCTS: A Simple Sampling Strategy for Data Synthesis (2025.acl-long)
Copied to clipboard
| Challenge: | Existing methods for generating multi-step reasoning data rely on rejection sampling, which generates trajectories independently and suffers from inefficiency and imbalanced sampling across problems of varying difficulty levels. |
| Approach: | They propose a data synthesis strategy inspired by Monte Carlo Tree Search . it offers step-level evaluation signals and promotes balanced sampling . |
| Outcome: | Experiments show that FastMCTS generates 30% more correct reasoning paths than rejection sampling. |
Unified Demonstration Retriever for In-Context Learning (2023.acl-long)
Copied to clipboard
Xiaonan Li, Kai Lv, Hang Yan, Tianyang Lin, Wei Zhu, Yuan Ni, Guotong Xie, Xiaoling Wang, Xipeng Qiu
| Challenge: | In-context learning is a new learning paradigm where a language model conditions on a few input-output pairs (demonstrations) and a test input, and directly outputs the prediction. |
| Approach: | They propose a single model to retrieve demonstrations for a wide range of tasks by combining training signals from various tasks into a unified list-wise ranking formulation by language model’s feedback. |
| Outcome: | The proposed model outperforms baselines on 30+ tasks across 13 task families and multiple data domains. |
Firewall Routing: Blocking Leads to Better Hybrid Inference for LLMs (2025.emnlp-main)
Copied to clipboard
| Challenge: | Large language models have significantly enhanced performance across various NLP tasks . high computational costs and latency associated with deploying such models pose bottlenecks . |
| Approach: | They propose a dynamic hybrid inference framework that efficiently selects between a strong and a weak LLM based on the complexity of the query. |
| Outcome: | The proposed method outperforms existing routing strategies by up to 5.29% in APGR . large models often introduce higher latency, making them less suitable for real-time or resource-constrained applications. |
LongWanjuan: Towards Systematic Measurement for Long Text Quality (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Existing efforts to improve data quality have focused on deduplication and the evaluation of data diversity and difficulty. |
| Approach: | They propose a set of metrics to evaluate the quality of long texts by evaluating three fundamental linguistic dimensions: coherence, cohesion, and complexity. |
| Outcome: | The proposed model improves on long-text tasks with over 160B tokens and categorizes long texts into holistic, aggregated, and chaotic types. |
Towards Transferable Personality Representation Learning based on Triplet Comparisons and Its Applications (2025.emnlp-main)
Copied to clipboard
Kai Tang, Rui Wang, Renyu Zhu, Minmin Lin, Xiao Ding, Tangjie Lv, Changjie Fan, Runze Wu, Haobo Wang
| Challenge: | Existing methods for personality analysis treat corpus as a single unit for classification, but this approach presents several challenges. |
| Approach: | They propose a task paradigm for text-based personality representation learning that uses a triplet personality trend comparison dataset to learn single-sentence personality embeddings with desirable metric properties. |
| Outcome: | The proposed model significantly boosts performance across various applications, including personality detection, personality retrieval, and emotion translation prediction. |
Adversarial Knowledge Stimulated Contrastive Prompting for Few-shot Language Learners (2023.findings-acl)
Copied to clipboard
| Challenge: | Prompt-based fine-tuning has boosted performance of Pre-trained language models on few-shot Natural Language Understanding (NLU) tasks by employing task-specific prompts. |
| Approach: | They propose a Cloze-driven prompt framework for prompt tuning that implicitly stimulates knowledge from pre-trained language models. |
| Outcome: | The proposed framework outperforms state-of-the-art for prompt-based fine-tuning on few-shot NLU tasks. |
Intuitive Fine-Tuning: Towards Simplifying Alignment into a Single Process (2025.acl-long)
Copied to clipboard
| Challenge: | Supervised Fine-Tuning (SFT) and Preference Optimization (PO) are key processes for aligning Language Models with human preferences post pre-training. |
| Approach: | They propose to combine Supervised Fine-Tuning and Preference Optimization (PO) with two sub-processes defined at token level within the Markov Decision Process (MDP) |
| Outcome: | The proposed process performs comparably or even superiorly to SFT and some typical PO methods across several tasks, particularly those requires generation, reasoning, and fact-following abilities. |
Full Parameter Fine-tuning for Large Language Models with Limited Resources (2024.acl-long)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) require massive GPU resources for training. |
| Approach: | They propose a parameter-efficient optimization that fuses the gradient computation and parameter update in one step to reduce memory usage. |
| Outcome: | The proposed method reduces memory usage to 10.8% compared to the standard approach. |
What are the Essential Factors in Crafting Effective Long Context Multi-Hop Instruction Datasets? Insights and Best Practices (2025.acl-long)
Copied to clipboard
| Challenge: | Existing methods to generate long-context instruction-tuning data are limited by poor quality and fewer than 35% of samples are multi-hop . |
| Approach: | They propose a framework that integrates a quality verification agent, a single-hop question generation agent, and a multi-hop questions merger agent to enhance model performance. |
| Outcome: | The proposed framework significantly improves data quality with high-quality, multi-hop, and diverse data. |
AdaLomo: Low-memory Optimization with Adaptive Learning Rate (2024.findings-acl)
Copied to clipboard
| Challenge: | Large language models require substantial memory for training, thereby setting a high hardware threshold. |
| Approach: | They propose a low-memory optimization technique that reduces memory footprint . they propose an adaptive learning rate for each parameter and a grouped update normalization to stabilize convergence . |
| Outcome: | The proposed low-memory optimization performs better than the prevailing algorithm for large language models, AdamW. |
CoLLiE: Collaborative Training of Large Language Models in an Efficient Way (2023.emnlp-demo)
Copied to clipboard
Kai Lv, Shuo Zhang, Tianle Gu, Shuhao Xing, Jiawei Hong, Keyu Chen, Xiaoran Liu, Yuqing Yang, Honglin Guo, Tengxiao Liu, Yu Sun, Qipeng Guo, Hang Yan, Xipeng Qiu
| Challenge: | Large language models (LLMs) are increasingly pivotal in a wide range of tasks . however, the resources required for training these models necessitate efficient solutions . |
| Approach: | They propose a library that facilitates collaborative training of large language models . they use 3D parallelism, parameter-efficient fine-tuning methods and optimizers . |
| Outcome: | The proposed library has proven superior training efficiency in comparison with prevalent solutions in pre-training and fine-tuning scenarios. |
ReAL: How Can LLMs Simulate the Real Teacher? Retrieval-enhanced Agent for Adaptive Learning (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Prior methods model learner-item interactions based only on ID sequences, leading to insufficient use of both learner and item information. |
| Approach: | They propose a Retrieval-enhanced Agent for Adaptive Learning powered by large language models to simulate teacher decision-making with extensive prior knowledge and teaching experience. |
| Outcome: | The proposed model outperforms existing models on three real-world datasets in both internal and external perspectives. |
CritiQ: Mining Data Quality Criteria from Human Preferences (2025.acl-long)
Copied to clipboard
Honglin Guo, Kai Lv, Qipeng Guo, Tianyi Liang, Zhiheng Xi, Demin Song, Qiuyinzhe Zhang, Yu Sun, Kai Chen, Xipeng Qiu, Tao Gui
| Challenge: | Existing methods to train language models rely on manual design, perplexity, or careful prompt engineering. |
| Approach: | They propose a method that automatically mines criteria from human preferences for data quality with only 30 human-annotated pairs and performs efficient data selection. |
| Outcome: | The proposed method improves on human-annotated test sets and shows high accuracy on code, math, and logic domains. |