Papers with stage training framework
Measuring and Improving Chain-of-Thought Reasoning in Vision-Language Models (2024.naacl-long)
Copied to clipboard
| Challenge: | Vision-language models have demonstrated strong efficacy as visual assistants . however, evaluation of their reasoning capabilities requires a costly benchmark . |
| Approach: | They propose a pipeline to measure the reasoning consistency of vision-language models . they propose supervised fine-tuning of VLMs and feedback from LLMs . |
| Outcome: | The proposed framework reduces cost while ensuring the generation of a high-quality dataset. |
Datasets and Recipes for Video Temporal Grounding via Reinforcement Learning (2025.emnlp-industry)
Copied to clipboard
Ruizhe Chen, Tianze Luo, Zhiting Fan, Heqing Zou, Zhaopeng Feng, Guiyang Xie, Hansheng Zhang, Zhuochen Wang, Zuozhu Liu, Zhang Huaijian
| Challenge: | Existing methods for video temporal grounding suffer from limited temporal awareness and poor generalization. |
| Approach: | They propose a two-stage training framework that integrates supervised fine-tuning with reinforcement learning to improve both the accuracy and robustness of VTG models. |
| Outcome: | The proposed training framework outperforms existing models on multiple benchmarks on open-domain and challenging scenarios. |
Unsupervised Single Document Abstractive Summarization using Semantic Units (2022.aacl-main)
Copied to clipboard
| Challenge: | a lack of sufficient training pairs is a common issue in real-world applications. |
| Approach: | They propose a framework that lets a model learn the frequency of each semantic unit in the source text. |
| Outcome: | The proposed model outperforms other unsupervised methods under CNN/Daily Mail task. |
Enhancing Neural Models with Vulnerability via Adversarial Attack (2020.coling-main)
Copied to clipboard
| Challenge: | Existing work on adversarial attack to improve performance of NLSM tasks has not been done. |
| Approach: | They propose a general two-stage training framework to enhance neural models with Vulnerability via adversarial attack. |
| Outcome: | The proposed framework improves neural models with Vulnerability via adversarial attack on NLSM datasets. |
Data Selection Curriculum for Neural Machine Translation (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Neural Machine Translation models are typically trained on heterogeneous data that are concatenated and randomly shuffled. |
| Approach: | They propose a two-stage curriculum training framework where a NMT model is fine-tuned on subsets of data, selected by deterministic scoring and online scoring. |
| Outcome: | The proposed framework improves on six language pairs comprising low- and high-resource languages and shows up to +2.2 BLEU improvement and faster convergence. |
Expanding the Boundaries of Vision Prior Knowledge in Multi-modal Large Language Models (2026.eacl-long)
Copied to clipboard
Qiao Liang, Yanjiang Liu, Weixiang Zhou, Ben He, Yaojie Lu, Hongyu Lin, Jia Zheng, Xianpei Han, Le Sun, Yingfei Sun
| Challenge: | Existing research treats MLLMs as unified systems optimized through end-to-end training, but the impact of vision encoder’s prior knowledge is seldom investigated. |
| Approach: | They propose a metric to quantify the effect of prior knowledge on MLLM performance by integrating prior knowledge at the vision encoder level into a training framework. |
| Outcome: | The proposed training framework incorporates prior knowledge at the vision encoder level, and significantly boosts visual understanding capabilities of MLLMs. |
D2Plan: Dual-Agent Dynamic Global Planning for Complex Retrieval-Augmented Reasoning (2026.acl-long)
Copied to clipboard
| Challenge: | Recent advances in reinforcement learning (RL) have empowered Large Language Models (LLMs) with the capability to perform autonomous retrieval during reasoning tasks. |
| Approach: | They propose a "D2Plan" paradigm for retrieval-augmented reasoning that integrates a 'Reasoner' and a'Purifier' |
| Outcome: | Experiments show that the proposed paradigm improves on QA benchmarks. |
Cross-Align: Modeling Deep Cross-lingual Interactions for Word Alignment (2022.emnlp-main)
Copied to clipboard
| Challenge: | Existing word alignment models capture few interactions between input sentence pairs, which severely degrades the word alignment quality. |
| Approach: | They propose to model deep interactions between input and target sentences using a two-stage training framework to train the model. |
| Outcome: | The proposed model achieves the state-of-the-art (SOTA) performance on four out of five language pairs. |
Relevance Is a Guiding Light: Relevance-aware Adaptive Learning for End-to-end Task-oriented Dialogue System (2024.emnlp-main)
Copied to clipboard
| Challenge: | Existing approaches to training task-oriented dialogue systems struggle with the Distractive Attributes Problem (DAP) Existing methods struggle to deal with false but similar knowledge (hard negative entities) |
| Approach: | They propose a two-stage training framework that eliminates hard negatives step-by-step and aligns retrieval with generation. |
| Outcome: | The proposed method eliminates hard negatives step-by-step and aligns retrieval with generation. |
Translatotron-V(ison): An End-to-End Model for In-Image Machine Translation (2024.findings-acl)
Copied to clipboard
| Challenge: | In-image machine translation (IIMT) aims to translate an image containing texts in source language into an image with translations in target language. |
| Approach: | They propose an end-to-end IIMT model with four modules that translate images . they propose a two-stage training framework to assist the model in learning alignment across languages . |
| Outcome: | The proposed model outperforms cascaded models with only 70.9% of parameters and is highly accurate. |
On Synthetic Data Strategies for Domain-Specific Generative Retrieval (2025.acl-long)
Copied to clipboard
| Challenge: | Generative retrieval models can be used to generate ranked lists of potentially relevant document identifiers for a user query. |
| Approach: | They propose a synthetic data generation strategy for a two-stage training framework that focuses on learning to decode document identifiers from queries and a strategy for mining hard negatives based on initial model's predictions. |
| Outcome: | The proposed model can generate ranked lists of potentially relevant document identifiers for a user query and then refine ranking through preference learning. |
Reward Modeling for Scientific Writing Evaluation (2026.acl-long)
Copied to clipboard
| Challenge: | Existing models for scientific writing evaluation are primarily optimized for general-purpose benchmarks with fixed scoring rubrics and evaluation criteria. |
| Approach: | They propose to train scientific writing evaluation models that leverage domain knowledge . they use a two-stage evaluation framework that optimizes evaluation preferences and refines reasoning capabilities . |
| Outcome: | The proposed model generalizes effectively across tasks and to previously unseen settings. |
Stimulate the Critical Thinking of LLMs via Debiasing Discussion (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing studies show that large language models (LLMs) are often prone to stance homogeneity and human preference biases when faced with conflicting perspectives. |
| Approach: | They propose a novel two-stage training framework to address stance homogeneity bias and human preference bias by generating multi-model discussion datasets and optimizing reinforcement learning from human feedback to align with discussion correctness. |
| Outcome: | The proposed framework reduces stance homogeneity bias and human preference bias and improves generalization capabilities on non-discussion scenarios and out-of-domain datasets. |
AFMRL: Attribute-Enhanced Fine-Grained Multi-Modal Representation Learning in E-commerce (2026.findings-acl)
Copied to clipboard
| Challenge: | Multimodal representation is crucial for E-commerce tasks such as identical product retrieval. |
| Approach: | They propose an approach which leverages the generative power of Multimodal Large Language Models to extract key attributes from product images and text and enhances representation learning through a two-stage training framework. |
| Outcome: | The proposed model achieves state-of-the-art on multiple downstream retrieval tasks, validating the effectiveness of harnessing generative models to advance fine-grained representation learning. |
Structured and Abstractive Reasoning on Multi-modal Relational Knowledge Images (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing studies on understanding and reasoning with abstractive information from the visual modality have not explored the use of STructured and Abstractive Reasoning (STAR) on such data. |
| Approach: | They propose an automatic STAR data engine to synthesize images with MMRK to build multi-modal instructions with reliable chain-of-thought thinking for various STAR tasks. |
| Outcome: | The proposed framework outperforms GPT-4o in STAR and improves performance across 8 open-source MLLMs. |
Improving the Robustness of Large Language Models via Consistency Alignment (2024.lrec-main)
Copied to clipboard
Yukun Zhao, Lingyong Yan, Weiwei Sun, Guoliang Xing, Shuaiqiang Wang, Chong Meng, Zhicong Cheng, Zhaochun Ren, Dawei Yin
| Challenge: | Large language models have shown tremendous success in following user instructions and generating helpful responses, but their robustness is still far from optimal. |
| Approach: | They propose a two-stage training framework that helps a model generalize on following instructions via similar instruction augmentations. |
| Outcome: | The proposed training framework improves diversity and aligns the model with human expectations by differentiating subtle differences in similar responses. |
VocalNet: Speech LLMs with Multi-Token Prediction for Faster and High-Quality Generation (2025.emnlp-main)
Copied to clipboard
| Challenge: | Experimental results show VocalNet outperforms existing open-source speech LLMs despite limited training data. |
| Approach: | They propose a scalable and model-agnostic training framework and a novel multi-token prediction paradigm for speech generation. |
| Outcome: | The proposed model outperforms open-source speech LLMs while outperforming existing open-sourced models. |
CompTab: A Comprehensive Benchmark for Real-World TableQA with Complex Reasoning and Irregular Tables (2026.acl-long)
Copied to clipboard
Zhen Yang, Wei Du, Jie Wang, Wenze Zhou, Xiangfeng Meng, Zhengyang Wang, Suping Sun, Ziwei Du, Haodong Zou, Jie Chen, Yongbin Liu, Shicheng Tan, Jiahao Ying, Shu Zhao
| Challenge: | Existing benchmarks focus on well-structured tables and fail to reflect irregular structures and complex reasoning commonly encountered in real-world scenarios. |
| Approach: | They propose a benchmark to evaluate TableQA under complex reasoning and irregular table conditions. |
| Outcome: | The proposed framework improves generalization and realism of large language models under complex and irregular table conditions. |
CARO: Chain-of-Analogy Reasoning Optimization for Robust Content Moderation (2026.findings-acl)
Copied to clipboard
| Challenge: | Current large language models struggle with ambiguous content moderation cases due to misleading "decision shortcuts" . authors propose a two-stage training framework to induce robust analogical reasoning in LLMs . |
| Approach: | They propose a two-stage training framework to induce robust analogical reasoning in LLMs . they bootstrap analogy reasoning chains via retrieval-augmented generation and SFT . |
| Outcome: | The proposed framework outperforms state-of-the-art reasoning models and specialized moderation models on ambiguous moderation benchmarks. |
RouteRAG: Efficient Retrieval-Augmented Generation from Text and Graph via Reinforcement Learning (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing graph-based or hybrid systems lack the ability to integrate supplementary evidence as reasoning unfolds. |
| Approach: | They propose a framework that integrates non-parametric knowledge into Large Language Models . they use a RL-based framework to optimize the entire generation process via RL . |
| Outcome: | The proposed framework outperforms existing RAG frameworks in five question answering benchmarks. |
LoRE: Enhancing Search Relevance with Progressive Chain-of-Thought and Preference Alignment (2026.findings-acl)
Copied to clipboard
Chenji Lu, Zhuo Chen, Hui Zhao, Zhiyuan Zeng, Gang Zhao, Junjie Ren, null Lihaoran, Songyan Liu, Pengjie Wang, Chuan Yu, Jian Xu, Bo Zheng
| Challenge: | E-commerce search relevance is a critical component of retrieval systems. |
| Approach: | They propose a large-generative model for search relevance that trains reasoning knowledge, multi-modal understanding and rule awareness into three core competencies. |
| Outcome: | The proposed model outperforms GPT-5 in Macro-F1 and achieves 27% online gain. |
Text Detoxification: Data Efficiency, Semantic Preservation and Model Generalization (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for detoxification of text often rely on manually annotated data . xiangli: "detoxification of texts is a powerful way to remove toxic content" |
| Approach: | They propose a reinforcement learning framework that optimizes detoxification and semantic preservation without annotating large amounts of data. |
| Outcome: | The proposed method overcomes major limitations and surpasses humanannotated references across multiple benchmarks. |
From Recognition to Reasoning: Benchmarking and Enhancing MLLMs on Real-World Receipt Document Understanding (2026.acl-long)
Copied to clipboard
| Challenge: | Existing models for visual information extraction suffer from limitations in scale and realism . ReceiptBench is a large-scale, human-annotated benchmark for receipts . |
| Approach: | They propose a large-scale, human-annotated benchmark for visual information extraction . the method organizes information extraction into four hierarchical sub-tasks . |
| Outcome: | The proposed method surpasses proprietary models on complex reasoning tasks. |