Papers by Chao Feng

31 papers
DisCo_Speech: Controllable Zero-Shot Speech Generation with A Disentangled Speech Codec (2026.acl-long)

Copied to clipboard

Challenge: DisCo-Speech is a zero-shot controllable text-to-speech framework . standard codecs entangle timbre and prosody, which hinders independent control in continuation-based LMs.
Approach: They propose a disentangled speech codec and an LM-based generator to solve this problem . they propose fusion and reconstruction that merges content and prosody into unified tokens .
Outcome: DisCo-Speech achieves competitive voice cloning and superior zero-shot prosody control.
Getting More from Less: Large Language Models are Good Spontaneous Multilingual Learners (2024.emnlp-main)

Copied to clipboard

Challenge: Large Language Models (LLMs) have shown impressive language capabilities, but most of them have very unbalanced performance across different languages.
Approach: They propose to use question translation data to enhance LLMs' multilingual capabilities by using mechanistic interpretability methods.
Outcome: The proposed method improves multilingual alignment even with unannotated answers in English and a wide range of languages even with instruction-tuned LLMs.
Doc2SoarGraph: Discrete Reasoning over Visually-Rich Table-Text Documents via Semantic-Oriented Hierarchical Graphs (2024.lrec-main)

Copied to clipboard

Challenge: Existing work on document visual question answering fails to capture the differences and correlations between elements of a document and associated questions.
Approach: They propose a document-visual question-answering challenge that exploits element-level semantics and employs hierarchical Graph structures to capture differences and correlations between elements.
Outcome: The proposed model surpasses the state-of-the-art method and large language model in terms of Exact Match (EM) metric, demonstrating exceptional effectiveness.
Boosting Multi-modal Keyphrase Prediction with Dynamic Chain-of-Thought in Vision-Language Models (2025.emnlp-main)

Copied to clipboard

Challenge: Multi-modal keyphrase prediction (MMKP) aims to produce concise, informative phrases that capture the essence of cross-modal inputs.
Approach: They propose to use vision-language models to generate conclusive phrases using multiple modalities of input information.
Outcome: The proposed methods outperform existing methods on absence and unseen scenarios and overestimate model capability due to overlap in training tests.
Spec-VLA: Speculative Decoding for Vision-Language-Action Models with Relaxed Acceptance (2025.emnlp-main)

Copied to clipboard

Challenge: Visual Language Models (VLMs) have significant parameter size and autoregressive (AR) decoding nature impose considerable computational demands on VLA models.
Approach: They propose a framework to relax acceptance utilizing the relative distances represented by the action tokens of the VLA model.
Outcome: Empirical results show that the proposed framework improves the speed of the prediction task by 44%.
Log-FGAER: Logic-Guided Fine-Grained Address Entity Recognition from Multi-Turn Spoken Dialogue (2023.emnlp-main)

Copied to clipboard

Challenge: Existing name entity recognition methods combine pre-trained language models with supervised models such as BiLSTM/LSTM-CRF to perform poorly in a spoken dialogue context.
Approach: They propose a logic-guided fine-grained address recognition method that softly applies the logic rule to improve the accuracy of FGAER.
Outcome: The proposed method improves fine-grained address entity recognition from multi-turn spoken dialogues.
Smarter, not Bigger: Fine-Tuned RAG-Enhanced LLMs for Automotive Hardware-in-the-Loop Testing (2026.acl-industry)

Copied to clipboard

Challenge: Hardware-in-the-Loop (HIL) testing is essential for automotive validation but suffers from fragmented and underutilized test artifacts.
Approach: They propose to integrate semantic retrieval with domain-adapted large language models to support test engineers in real-world HIL workflows.
Outcome: The proposed system improves perceived helpfulness, truthfulness, and satisfaction over general-purpose LLMs.
Rewrite to Jailbreak: Discover Learnable and Transferable Implicit Harmfulness Instruction (2025.findings-acl)

Copied to clipboard

Challenge: Existing jailbreak methods create a forced instruction-following scenario, or search adversarial prompts with prefix or suffix tokens to achieve a specific representation manually or automatically.
Approach: They propose a method that rewrites the original instruction to achieve a jailbreak . they propose rewriting the original instructions to improve the attack strategy .
Outcome: The proposed method is more efficient and easier to identify since no additional features are introduced.
Reinforced Product Metadata Selection for Helpfulness Assessment of Customer Reviews (D19-1)

Copied to clipboard

Challenge: a helpful review is largely concerned with the metadata of its target product . a selector learns from both the key-value product metadata and one of its reviews to take an action .
Approach: They propose a framework that uses product metadata to assess helpfulness of free-text reviews . they use two real-world datasets from amazon.com and Yelp.com to test the framework .
Outcome: The proposed framework can achieve state-of-the-art performance with substantial improvements . it uses two real-world datasets from Amazon.com and Yelp.com .
Stumbling Blocks: Stress Testing the Robustness of Machine-Generated Text Detectors Under Attacks (2024.acl-long)

Copied to clipboard

Challenge: Existing studies on this topic focus on the robustness of specific detectors or particular attack methods.
Approach: They stress test the detectors’ robustness to malicious attacks under realistic scenarios using LLMs and metric-based detectors.
Outcome: The proposed methods are based on a set of LLM-based models and their performance is compared under different budget levels.
TAT-QA: A Question Answering Benchmark on a Hybrid of Tabular and Textual Content in Finance (2021.acl-long)

Copied to clipboard

Challenge: Existing QA systems focus on unstructured text, structured knowledge base, or semi-structured tables.
Approach: They propose a large-scale question answering model based on financial reports . numerical reasoning is usually required to infer the answer .
Outcome: The proposed model achieves 58.0% inF1, an 11.1% increase over the baseline model, but still lags behind the best human model.
CERES: Pretraining of Graph-Conditioned Transformer for Semi-Structured Session Data (2022.naacl-main)

Copied to clipboard

Challenge: Despite advances in self-supervised learning, there is a lack of models that can effectively capture both intra- and intra-item semantics for semi-structured session data.
Approach: They propose a graph-based transformer model for semi-structured session data that captures both intra- and intra-item semantics.
Outcome: The proposed model outperforms baselines in three session search and entity linking tasks by up to 9%.
Attribution-Based Analysis and Optimization of Modular Agentic Workflows (2026.findings-acl)

Copied to clipboard

Challenge: Large Language Models (LLMs) have driven the rise of agentic workflows . yet, how can we attribute performance gains to individual upgrades and their interactions?
Approach: They propose a game-theoretic framework that models component upgrades as players and evaluates component coalitions to compute Shapley values.
Outcome: The proposed framework provides interaction-aware attribution and recommendation for model allocation under a fixed workflow structure.
Investigating and Scaling up Code-Switching for Multilingual Language Model Pre-Training (2025.findings-acl)

Copied to clipboard

Challenge: Large language models (LLMs) exhibit remarkable multilingual capabilities despite the extreme language imbalance in the pre-training data.
Approach: They investigate the existence of code-switching in the pre-training corpus and categorize it into four types within two quadrants.
Outcome: The proposed approach improves performance across benchmarks and representation space.
MultiPL-MoE: Multi-Programming-Lingual Extension of Large Language Models through Hybrid Mixture-of-Experts (2025.findings-emnlp)

Copied to clipboard

Challenge: MultiPL is a special case of multiple natural languages and requires limited computational resources to generate multilingual code.
Approach: They propose to extend LLMs by combining two paired experts to optimize expert selection at token and segment levels.
Outcome: The proposed extension improves the performance of the base LLMs while retaining the most popular ones using limited computational resources.
Speech-based Slot Filling using Large Language Models (2024.findings-acl)

Copied to clipboard

Challenge: Recent advances in large language models (LLMs) have shown an unprecedented ability across various language tasks.
Approach: They propose to use prompts and LoRA fine-tuning to improve slot filling robustness . they propose a linearised knowledge injection scheme to integrate dynamic external knowledge into LLMs.
Outcome: The proposed model improves slot filling with noisy ASR transcriptions with 6.7% and 17.6% absolute SLU-F1 improvements compared to a fully fine-tuned Flan-T5-XL model.
Large Language Models Are Cross-Lingual Knowledge-Free Reasoners (2025.naacl-long)

Copied to clipboard

Challenge: Large language models have demonstrated impressive reasoning capabilities across multiple languages, but the relationship between capabilities in different languages is less explored.
Approach: They decompose the process of reasoning tasks into two separate components: knowledge retrieval and knowledge-free reasoning.
Outcome: The proposed model can be transferred across source-target languages despite secondary impact of resource in some specific target languages, while cross-lingual knowledge retrieval significantly hinders the transfer.
Palette of Language Models: A Solver for Controlled Text Generation (2025.naacl-long)

Copied to clipboard

Challenge: Recent advances in large language models have revolutionized text generation with their remarkable capabilities.
Approach: They propose to combine a single-attribute model with a discriminative model to achieve a combination strategy that incorporates positive correlation and attribute enhancement.
Outcome: The proposed method is adapted for single-attribute control scenario and achieves surpassing results.
Guided by the Plan: Enhancing Faithful Autoregressive Text-to-Audio Generation with Guided Decoding (2026.eacl-long)

Copied to clipboard

Challenge: Autoregressive (AR) models excel at generating temporally coherent audio by producing tokens sequentially, yet they often falter in faithfully following complex textual prompts.
Approach: They propose a lightweight auxiliary model trained with a GAE-inspired objective to predict final instruction-following quality from partial generations.
Outcome: The proposed model achieves 10 points improvement in CLAP score over baseline AR models while maintaining computational parity with best-of-N decoding.
PLaD: Preference-based Large Language Model Distillation with Pseudo-Preference Pairs (2024.findings-acl)

Copied to clipboard

Challenge: Knowledge distillation (KD) is a technique for transferring expertise from large teacher models to compact student models with reduced memory footprints and inference costs.
Approach: They propose to transfer knowledge from large teacher models to compact student models by exploiting teacher-student capacity discrepancies to generate pseudo-preference pairs where teacher outputs are preferred over student outputs.
Outcome: The proposed framework exploits teacher-student capacity discrepancy to generate pseudo-preference pairs where teacher outputs are preferred over student outputs.
Masked Diffusion Captioning for Visual Feature Learning (2025.findings-emnlp)

Copied to clipboard

Challenge: Unlike autoregressive captioning, the strength of the visual learning signal in MDC does not depend on each token’s position in the sequence, reducing the need for auxiliary objectives.
Approach: a decoder conditioned on visual features is trained to reconstruct the original text.
Outcome: masked diffusion captioning (MDC) is a form of image-conditioned captioning that can be applied to visual tasks.
Chain-of-Procedure: Hierarchical Visual-Language Reasoning for Procedural QA (2026.findings-acl)

Copied to clipboard

Challenge: Recent advances in vision-language models (VLMs) have achieved impressive results on standard image-text tasks, yet their capability in visual procedure question answering (VP-QA) remains largely unexplored.
Approach: They propose a multimodal benchmark specifically designed for visual procedural reasoning that synergizes cross-modal procedure retrieval, context-aware step decomposition, and the next step prediction.
Outcome: The proposed framework significantly outperforms baselines on visual procedure question answering (VP-QA) Experiments on six VLMs show that it performs better than baselines.
Lost in the Context: Insufficient and Distracted Attention to Contexts in Preference Modeling (2025.acl-long)

Copied to clipboard

Challenge: Existing reward models concatenate contexts and responses, but they often ignore crucial segments of the context that are important for evaluating the response quality.
Approach: They propose a reward model that evaluates the response quality based on a given context and assigns a rewards reward.
Outcome: The proposed framework significantly improves preference modeling by increasing attention to relevant information within the context and achieves better generalizability.
LLM as a metric critic for low resource relation identification (2024.findings-emnlp)

Copied to clipboard

Challenge: Existing studies show that small language models (SLMs) overfit in low resource situations . however, the gap between pre-training and fine-tuning leads to performance decay .
Approach: They propose to combine large language models and LLM for relation identification by co-evolution . they propose to use a masked language model prompt to generate a relation identification task .
Outcome: The proposed model can handle low resource relation identification tasks with minimal overfitting . the proposed model provides essential background knowledge to assist training process .
Vision-Flan: Scaling Human-Labeled Tasks in Visual Instruction Tuning (2024.findings-acl)

Copied to clipboard

Challenge: Recent vision-language models (VLMs) have shown impressive capabilities as general visual assistants, but there are two challenges to their performance: (1) lacking task diversity in pretraining and visual instruction tuning; (2) annotation error and bias in GPT-4 synthesized instruction tuning data.
Approach: They propose a two-stage instruction tuning framework that fine tunes VLMs firstly and further tuned on GPT-4 synthesized data.
Outcome: The proposed framework outperforms the traditional single-stage visual instruction tuning framework and achieves state-of-the-art performance across a wide range of multi-modal evaluation benchmarks.
Understanding LLMs’ Cross-Lingual Context Retrieval: How Good It Is And Where It Comes From (2025.emnlp-main)

Copied to clipboard

Challenge: Cross-lingual context retrieval is a fundamental aspect of cross-lingual alignment, but the performance and mechanism of it for large language models (LLMs) remains unclear.
Approach: They evaluate cross-lingual context retrieval of over 40 large language models . they use cross-linguistic machine reading comprehension as a representative scenario .
Outcome: The results show that open LLMs show strong cross-lingual context retrieval ability . the results also show that their oracle performances improve after training .
Trust Within? Seek Beyond? Knowledge Boundary Aware Policy Optimization for Agentic Search (2026.acl-long)

Copied to clipboard

Challenge: Existing approaches to augment large language models with external knowledge suffer from a lack of calibration regarding the model’s knowledge boundary.
Approach: They propose a reinforcement learning framework that explicitly aligns retrieval decisions with quantified knowledge states.
Outcome: The proposed framework outperforms strong baselines while exhibiting reduced hallucination rates.
Diversity and Consistency: Exploring Visual Question-Answer Pair Generation (2021.findings-emnlp)

Copied to clipboard

Challenge: Existing tasks to generate question-answer pairs from visual images are under-explored.
Approach: They propose a task that targets question-answer pair generation from visual images.
Outcome: The proposed model can generate diverse or consistent QAPs on two benchmarks.
Self-attention-based Graph-of-Thought for Math Problem Solving (2025.findings-acl)

Copied to clipboard

Challenge: Existing methods for generating reasoning paths in a chain structure are inefficient and non-human-like.
Approach: They propose a decoding method for a chain-based LLM that constructs a thought graph simultaneously as an LLM inference and generates reasoning steps with a graph-structured self-attention mechanism.
Outcome: The proposed method improves reasoning accuracy without huge computational over-expensive LLMs and avoids performance degradation issues when the LLM is too small to comprehend complex prompts.
Minimal, Local, and Robust: Embedding-Only Edits for Implicit Bias in T2I Models (2025.emnlp-main)

Copied to clipboard

Challenge: EmbEdit is a text-to-image editing method that only fine-tunes the word token embedding (WTE) of the target object.
Approach: They propose a method to edit implicit assumptions and priors in text-to-image models without affecting unrelated objects or degrading overall performance.
Outcome: The proposed method outperforms previous methods in various models, tasks, and editing scenarios.
Beyond Layout Embedding: Layout Attention with Gaussian Biases for Structured Document Understanding (2023.findings-emnlp)

Copied to clipboard

Challenge: Existing methods for encoding layout information rely on millions of learnable parameters . polar coordinates provide superior choice for layout modeling, study finds .
Approach: They propose to model layout attention with Gaussian biases by feeding polar coordinates into 2-D Gausssian kernels.
Outcome: The proposed model improves on three widely used benchmarks.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations