Papers by Yi Ma

35 papers
Symbol tuning improves in-context learning in language models (2023.emnlp-main)

Copied to clipboard

Challenge: Language models are sensitive to the way that prompts are given, indicating that they are not reasoning in a robust manner.
Approach: They propose to fine tune language models on in-context input-label pairs where natural language labels are replaced with arbitrary symbols.
Outcome: The proposed model is much stronger at reasoning tasks and more robust to underspecified prompts than the standard model.
MS2SL: Multimodal Spoken Data-Driven Continuous Sign Language Production (2024.findings-acl)

Copied to clipboard

Challenge: Existing solutions for sign language production are limited due to phonological differences and data scarcity.
Approach: They propose a unified framework for continuous sign language production that generates sign predictions step by step from text or speech embeddings.
Outcome: The proposed model achieves competitive performance on how2sign and PHOENIX14T datasets.
TextMixer: Mixing Multiple Inputs for Privacy-Preserving Inference (2023.findings-emnlp)

Copied to clipboard

Challenge: Pre-trained language models (PLMs) are often deployed as cloud services, enabling users to upload textual data and perform inference remotely.
Approach: They propose a privacy-preserving inference framework called MixPi which aims to obfuscate a user's private input by mixing it with multiple other inputs.
Outcome: The proposed framework surpasses existing privacy-preserving methods on token and sentence classification tasks.
Reasoning with Language Model is Planning with World Model (2023.emnlp-main)

Copied to clipboard

Challenge: Large language models (LLMs) have shown remarkable reasoning capabilities, particularly with Chain-of-Thought-style prompts.
Approach: They propose a framework that repurposes the LLM as both a world model and a reasoning agent and incorporates a principled planning algorithm (based on Monte Carlo Tree Search)
Outcome: The proposed framework repurposes the LLM as both a world model and a reasoning agent and incorporates a principled planning algorithm (based on Monte Carlo Tree Search) it achieves optimum balance between exploration and exploitation, while achieving high-reward reasoning paths efficiently.
PFDial: A Structured Dialogue Instruction Fine-tuning Method Based on UML Flowcharts (2025.findings-acl)

Copied to clipboard

Challenge: Large Language Models (LLMs) have shown remarkable progress in dialogue and reasoning, but they struggle to solve strictly constrained dialogue tasks.
Approach: They construct a dataset that contains 12,705 high-quality Chinese dialogue instructions from 440 flowcharts containing 5,055 process nodes.
Outcome: The proposed model outperforms GPT-4o models on backward transitions and outperformed GPT-42 models on the same dataset.
[MASK] Insertion: a robust method for anti-adversarial attacks (2023.findings-eacl)

Copied to clipboard

Challenge: Existing studies have focused on adversarial defenses against pretrained language models.
Approach: They propose an adversarial defensing algorithm that inserts tokens into input sequences . they show an improvement in accuracy between 3.2 and 11.1 absolute points .
Outcome: The proposed algorithm improves model accuracy on clean and polluted inputs compared with state-of-the-art models .
ThinkLinker: From Low-Rank Interaction to Knowledge-Aware Verification for Multimodal Entity Linking (2026.findings-acl)

Copied to clipboard

Challenge: Existing methods for multimodal entity linking rely on textual context for disambiguation . textual contextual information alone fails to resolve ambiguity, leading to unreliable disambiguations in weak contexts.
Approach: They propose a two-stage multimodal entity linking framework called ThinkLinker . they propose fusion mechanism to model joint dependencies among features .
Outcome: The proposed framework outperforms state-of-the-art models on public benchmark datasets.
PizzaPal: Conversational Pizza Ordering using a High-Density Conversational AI Platform (D18-2)

Copied to clipboard

Challenge: a pizza ordering bot that can be used to order pizzas is described in this paper.
Approach: They describe PizzaPal, a voice-only agent for ordering pizza, and the Conversational AI architecture built at b4.ai.
Outcome: The pizza ordering bot is based on a dialog framework developed by b4.ai .
M-ABSA: A Multilingual Dataset for Aspect-Based Sentiment Analysis (2025.emnlp-main)

Copied to clipboard

Challenge: Existing studies focus on English-centric aspects of sentiment analysis, limiting scope for multilingual evaluation and research.
Approach: They propose to use a multilingual dataset to analyze aspects with associated sentiment elements in text.
Outcome: The proposed dataset is the most extensive multilingual parallel dataset for ABSA to date.
Mixture of Heterogeneous Grouped Experts for Language Modeling (2026.acl-industry)

Copied to clipboard

Challenge: Large Language Models (LLMs) based on Mixture-of-Experts (MoE) enforce uniform expert sizes, creating a rigidity that fails to align computational costs with varying token-level complexity.
Approach: They propose a mixture of heterogeneous grouped experts (MoHGE) that allows for flexible, resource-aware expert combinations.
Outcome: The proposed model matches the performance of existing Mixture-of-Experts architectures while maintaining balanced GPU utilization.
PLAWBENCH: A Rubric-Based Benchmark for Evaluating LLMs in Real-World Legal Practice (2026.acl-long)

Copied to clipboard

Challenge: Existing benchmarks for large language models (LLMs) are coarse, single-dimensional metrics and do not explicitly assess fine-grained legal reasoning.
Approach: They propose a Practical Law Benchmark to evaluate large language models in real-world legal practice scenarios.
Outcome: The proposed model is based on 850 questions and 13 scenarios with expert-designed evaluation rubrics.
Making Pre-trained Language Models Better Continual Few-Shot Relation Extractors (2024.lrec-main)

Copied to clipboard

Challenge: Existing methods to learn new relations with limited labeled data are prone to catastrophic forgetting and overfitting.
Approach: They propose a framework that uses prompts to acquire more generalized knowledge . they propose CFRE to continuously learn new relations while retaining knowledge of old ones .
Outcome: The proposed method outperforms state-of-the-art methods by a large margin and significantly mitigates catastrophic forgetting and overfitting in low-resource scenarios.
Too Good to be Bad: On the Failure of LLMs to Role-Play Villains (2026.findings-acl)

Copied to clipboard

Challenge: Large Language Models (LLMs) are increasingly tasked with creative generation, but their ability to portray non-prosocial, antagonistic personas remains largely unexamined.
Approach: They propose a moral alignment benchmark to test the safety of large language models . they find that models struggle with traits directly antithetical to safety principles .
Outcome: The proposed model fails to accurately portray morally ambiguous or villainous characters . the model fails most with traits directly antithetical to safety principles .
WebAggregator: Enhancing Compositional Reasoning Capabilities of Deep Research Agent Foundation Models (2026.acl-long)

Copied to clipboard

Challenge: Existing agentic systems are retrieval-heavy but reasoning-light . current systems lack compositional reasoning, a key component of deep research .
Approach: They propose a data synthesis pipeline WebAggregator to shift agentic paradigm . they use Proactive Explorer to collect interconnected knowledge and Compositional Logic Proposer to weave knowledge into complex questions .
Outcome: The proposed pipeline surpasses GPT-4.1 and matches Claude-3.7-Sonnet on GAIA, WebWalkerQA, and XBench.
HiTRANS: A Hierarchical Transformer Network for Nested Named Entity Recognition (2021.findings-emnlp)

Copied to clipboard

Challenge: Existing studies for named entity recognition focus on flat NER, i.e., without nested entities, by sequence labeling methods.
Approach: They propose a Hierarchical Transformer network which decomposes the input sentence into multi-grained spans and enhances the representation learning in a hierarchical manner.
Outcome: The proposed method achieves much better performance than the state-of-the-art approaches on GENIA, ACE-2004, ace-2005 and NNE datasets.
Meta-Reinforced Multi-Domain State Generator for Dialogue Systems (2020.acl-main)

Copied to clipboard

Challenge: Existing methods to train a multi-domain dialogue state tracker are lacking in accuracy.
Approach: They propose a Meta-Reinforced Multi-Domain State Generator to train a DST meta-learning model with a few domains as source domains and a new domain as target domain.
Outcome: The proposed system outperforms the traditional training approach with extremely little training data in target domain.
A Coarse-to-fine Cascaded Evidence-Distillation Neural Network for Explainable Fake News Detection (2022.coling-1)

Copied to clipboard

Challenge: Existing methods for fake news detection focus on fact-checked reports, resulting in limited coverage and debunking delays.
Approach: They propose a Coarse-to-fine Cascaded Evidence-Distillation neural network for explainable fake news detection based on raw reports . they use hierarchical encoders and cascaded selectors to select most explainable sentences for verdicts on top of selected top-K reports based upon raw reports.
Outcome: The proposed model outperforms baseline detection methods and generates high-quality explanations from diverse evaluation perspectives.
COVID-19 Literature Knowledge Graph Construction and Drug Repurposing Report Generation (2021.naacl-demos)

Copied to clipboard

Challenge: a new framework to digest relevant biomedical knowledge is needed to combat COVID-19 . quantity of research results is a bottleneck, and false information promoted in publications .
Approach: a team of researchers has developed a framework to extract multimedia knowledge elements from scientific literature to combat COVID-19.
Outcome: a new framework extracts fine-grained multimedia knowledge elements from scientific literature . it provides detailed contextual sentences, subfigures, and knowledge subgraphs as evidence . the framework is based on a case study of drug repurposing .
Natural-Language Policies to Executable Decisions: An Interpretable Large Language Model Framework (2026.acl-industry)

Copied to clipboard

Challenge: a production-grade pricing system for tourism is challenging due to unstructured nature of travel orders and ever-evolving pricing policies.
Approach: They propose a production-grade pricing system with a strict decision boundary . they propose to combine structured extraction and bounded policy/path selection with interpretable condition trees .
Outcome: The proposed system processed 3,960 orders in six months and reduced the order management team from 15-20 to 3 . the system reduced the per-order handling time from 10 minutes to 2 minutes.
R3: End-to-End Reasoning-based Planning for Multi-step Retrosynthesis via Reinforcement Learning (2026.acl-long)

Copied to clipboard

Challenge: Experimental results show that R3 is a superior alternative to traditional search algorithms for multistep retrosynthesis planning.
Approach: They propose a framework that reformulates multistep retrosynthetic planning as a generative reasoning task.
Outcome: The proposed framework achieves state-of-the-art Top-1 accuracy of 43.7% on retrobench . it leverages Large Language Models to reformulate multistep retrosynthesis as a generative reasoning task.
ChatMusician: Understanding and Generating Music Intrinsically with LLM (2024.findings-acl)

Copied to clipboard

Challenge: Despite LLMs' impressive capabilities in musical knowledge, music reasoning remains an unsolved task.
Approach: They propose an open-source large language model (LLM) that integrates intrinsic musical abilities into LLaMA2 and GPT-3.5.
Outcome: The proposed model can understand and generate music with a pure text tokenizer without external multi-modal neural structures or tokenizers.
Towards Low-Resource Semi-Supervised Dialogue Generation with Meta-Learning (2020.findings-emnlp)

Copied to clipboard

Challenge: Existing systems that use labelled data to generate dialogues are lacking in high accuracy.
Approach: They propose a meta-learning based semi-supervised explicit dialogue state tracker for neural dialogue generation, denoted as MEDST.
Outcome: The proposed system outperforms existing systems by 18.7% goal accuracy and 14.3% entity match rate on the KVRET corpus with 2% labelled data in semi-supervision.
Large Dual Encoders Are Generalizable Retrievers (2022.emnlp-main)

Copied to clipboard

Challenge: Experimental results show that dual encoders outperform sparse and dense retrievers on the BEIR dataset significantly.
Approach: They challenge belief that bottleneck layer is too limited for out-of-domain generalization . they scale up the model while keeping bottleneck as a single dot-product with a fixed size .
Outcome: The proposed model outperforms sparse and dense retrievers on the BEIR dataset significantly.
CLIO: Role-interactive Multi-event Head Attention Network for Document-level Event Extraction (2022.coling-1)

Copied to clipboard

Challenge: Existing methods for document-level event extraction struggle due to two intrinsic challenges: nested arguments and multiple events.
Approach: They propose a role-interactive multi-event head attention network to solve two challenges . they map different events to multiple subspaces and then determine whether the current event exists .
Outcome: The proposed model improves on two widely used DEE datasets on the Internet.
UniCorn: Towards Self-Improving Unified Multimodal Models through Self-Generated Supervision (2026.acl-long)

Copied to clipboard

Challenge: Unified Multimodal Models have achieved remarkable success in cross-modal comprehension, but a gap persists in their ability to translate internal knowledge into faithful and controllable synthesis.
Approach: They propose a self-improvement framework that partitions a single UMM into three collaborative roles: Proposer, Solver, and Judge.
Outcome: The proposed framework improves on TIIF, DPG, CompBench and UniCycle benchmarks.
Domain-Specific Data Generation Framework for RAG Adaptation (2026.findings-acl)

Copied to clipboard

Challenge: Retrieval-Augmented Generation (RAG) combines the language understanding and reasoning capabilities of large language models (LLMs) with external retrieval to produce domain-grounded responses.
Approach: They propose a scalable and modular data-centric framework for generating domain-grounded question–answer–context triples tailored to diverse RAG adaptation strategies.
Outcome: The proposed framework generates domain-grounded question–answer–context triples for multiple RAG adaptation strategies.
PILE: Pairwise Iterative Logits Ensemble for Multi-Teacher Labeled Distillation (2022.emnlp-industry)

Copied to clipboard

Challenge: Pre-trained language models have been a key part of ranking systems . knowledge distillation is widely used to maintain high performance while keeping efficient computations.
Approach: They propose an algorithm to combine knowledge from multi-teachers and label information to achieve competitive performance in offline and online experiments.
Outcome: The proposed method has been deployed in a real-world commercial search system.
Stratagem: Learning Transferable Reasoning via Trajectory-Modulated Game Self-Play (2026.acl-long)

Copied to clipboard

Challenge: Existing self-play approaches to developing general reasoning in language models rely on terminal game outcomes.
Approach: They propose a game-based reasoning transfer model that addresses two barriers to reasoning transfer.
Outcome: The proposed model improves mathematical reasoning, general reasoning, and code generation benchmarks.
Making Harmful Behaviors Unlearnable for Large Language Models (2024.findings-acl)

Copied to clipboard

Challenge: Large language models (LLMs) are often customized by fine-tuning for the requirements of different domains.
Approach: They propose a controllable training framework to make undesired behaviors unlearnable during the fine-tuning process.
Outcome: The proposed framework makes undesired behaviors unlearnable during the fine-tuning process while preserving the ability to learn other information.
SLIM: Subtrajectory-Level Elimination for More Effective Reasoning (2025.findings-emnlp)

Copied to clipboard

Challenge: Notable examples include OpenAI’s o1/o3/o4 series and DeepSeek-R1 .
Approach: They develop a framework to identify suboptimal subtrajectories based on human-established criteria . they also use a sampling algorithm to select data whose reasoning process is free from suboptimally subtravertories to the highest degree .
Outcome: The proposed method reduces the number of suboptimal subtrajectories by 25.9% during the inference process.
Does Higher Order LSTM Have Better Accuracy for Segmenting and Labeling Sequence Data? (C18-1)

Copied to clipboard

Challenge: Existing neural models take long distance dependencies into account when predicting the tag of the current token.
Approach: They propose a method to capture long distance tag dependencies and use them for dependency analysis.
Outcome: The proposed model can predict multiple tags for the current token without taking dependencies between tags into account.
GraphICL: Unlocking Graph Learning Potential in LLMs through Structured Prompt Design (2025.findings-naacl)

Copied to clipboard

Challenge: Text-Attributed Graphs (TAGs) are a powerful tool for understanding complex systems and relationships.
Approach: They propose a benchmark to evaluate large language models for graph-structured data using prompts.
Outcome: The proposed benchmark outperforms state-of-the-art graph LLMs and graph neural networks on graph learning tasks without training.
SAVOIR: Learning Social Savoir-Faire via Shapley-based Reward Attribution (2026.findings-acl)

Copied to clipboard

Challenge: Existing approaches to improve social intelligence of AI systems employ retrospective attributions and lack theoretical grounding.
Approach: They propose a framework that uses Shapley values to ensure fair credit distribution with axiomatic guarantees of efficiency, symmetry, and marginality.
Outcome: The proposed framework matches or exceeds proprietary models including GPT-4o and Claude-3.5-Sonnet.
ED2LM: Encoder-Decoder to Language Model for Faster Document Re-ranking Inference (2022.findings-acl)

Copied to clipboard

Challenge: State-of-the-art neural models typically encode document-query pairs using cross-attention for re-ranking.
Approach: They propose to fine tune a pretrained encoder-decoder model using document to query generation.
Outcome: The proposed model achieves comparable results to more expensive approaches while being 6.8X faster.
TextObfuscator: Making Pre-trained Language Model a Privacy Protector via Obfuscating Word Representations (2023.findings-acl)

Copied to clipboard

Challenge: Existing inference services are plagued by privacy concerns, such as sharing sensitive data with service providers.
Approach: They propose a framework for protecting inference privacy by applying random perturbations to clustered representations.
Outcome: The proposed framework protects inference privacy by applying random perturbations to clustered representations.

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