Papers by Hong Sun

25 papers
Enhancing Self-Attention with Knowledge-Assisted Attention Maps (2022.naacl-main)

Copied to clipboard

Challenge: Existing works of knowledge infusion depend on multi-task learning frameworks, which are inefficient and require large-scale retraining when new knowledge is considered.
Approach: They propose a method which integrates knowledge-generated attention maps into the self-attention mechanism and integrates it into the model.
Outcome: The proposed model outperforms existing methods on academic datasets and industry-scale ad relevance applications.
OpenHuEval: Evaluating Large Language Model on Hungarian Specifics (2025.findings-acl)

Copied to clipboard

Challenge: Recent advances in Large Language Models (LLMs) represent significant strides toward artificial general intelligence (AGI).
Approach: They introduce OpenHuEval, the first benchmark for LLMs focusing on the Hungarian language and specifics.
Outcome: The framework reveals intrinsic patterns and mechanisms of LLMs in non-English languages, with Hungarian serving as an example.
Uncertainty Guided Label Denoising for Document-level Distant Relation Extraction (2023.acl-long)

Copied to clipboard

Challenge: Document-level relation extraction (DocRE) aims to extract semantic relations between entities in a document.
Approach: They propose a Document-level distant relation extraction framework with unreliable pseudo labels to denoise DS data.
Outcome: The proposed framework outperforms strong baselines on two public datasets.
Question Generation from SQL Queries Improves Neural Semantic Parsing (D18-1)

Copied to clipboard

Challenge: Using question generation, we learn a semantic parser with 30% of the supervised training data.
Approach: They propose to use question generation to learn a semantic parser with less supervised training data.
Outcome: The proposed method improves the state-of-the-art model with less training data.
Large Language Model Is Not a Good Few-shot Information Extractor, but a Good Reranker for Hard Samples! (2023.findings-emnlp)

Copied to clipboard

Challenge: Large Language Models have made remarkable strides in various tasks, but whether they are competitive few-shot solvers remains an open question.
Approach: They propose an adaptive filter-then-rerank paradigm to combine the strengths of LLMs and SLMs.
Outcome: The proposed system achieves promising improvements on various IE tasks with acceptable time and cost investment.
DischargeSim: A Simulation Benchmark for Educational Doctor–Patient Communication at Discharge (2025.emnlp-main)

Copied to clipboard

Challenge: Discharge communication is a critical yet underexplored component of patient care, where the goal shifts from diagnosis to education.
Approach: They propose a benchmark that evaluates large language models’ ability to act as personalized discharge educators.
Outcome: Experiments with 18 LLMs show that model size does not always yield better education outcomes, highlighting trade-offs in strategy use and content prioritization.
SOLAR: Serendipity Optimized Language Model Aligned for Recommendation (2025.findings-emnlp)

Copied to clipboard

Challenge: Large Language Models have shown strong potential in recommendation tasks . however, their application to serendipity-oriented recommendations remains challenging .
Approach: They propose a domain-adaptive instruction tuning method that aligns Large Language Models with recommendation tasks.
Outcome: The proposed framework bridges the domain gap between LLMs and recommendation tasks.
BBTv2: Towards a Gradient-Free Future with Large Language Models (2022.emnlp-main)

Copied to clipboard

Challenge: Recent work on parameter-efficient tuning (PET) only tunes a small portion of parameters while keeping most of the parameters of the LLM unchanged.
Approach: They propose an improved version of Black-Box Tuning to tune PTMs through gradient descent . they prepend continuous prompts to every layer of the PTM and propose a divide-and-conquer gradient-free algorithm to optimize the prompts alternately.
Outcome: The proposed method achieves comparable performance to full model tuning and state-of-the-art parameter-efficient methods under few-shot settings while maintaining much fewer tunable parameters.
Adversarial Alignment with Anchor Dragging Drift (A3D2): Multimodal Domain Adaptation with Partially Shifted Modalities (2025.acl-long)

Copied to clipboard

Challenge: Domain adaptation is underexplored in multimodal learning environments due to expensive data collection and annotation.
Approach: They propose a bi-alignment scheme to perform drift-drift and anchor-driving matching with partially shifting anchors.
Outcome: The proposed approach achieves superior performance compared with state-of-the-art approaches.
Using Subtext to Enhance Generative IDRR (2025.acl-short)

Copied to clipboard

Challenge: Arguments contain subtexts, but they are connotative and need prompts to be recognized . a lightweight subtext generator is helpful when the prompt doesn't raise a complex CoT.
Approach: They leverage LLaMA to generate subtexts for argument pairs and verify their effectiveness . they construct a baseline IDRR using the decoder-only backbone LLama .
Outcome: The proposed approach achieves higher F1 scores on two benchmarks than previous models.
Emma-X: An Embodied Multimodal Action Model with Grounded Chain of Thought and Look-ahead Spatial Reasoning (2025.acl-long)

Copied to clipboard

Challenge: Visual-Language-Action models lack the ability to generate actionable policies tailored to specific robotic embodiments.
Approach: They propose an embodied multimodal action model with Grounded Chain of Thought and Look-ahead Spatial Reasoning that enhances spatial reasoning and task planning.
Outcome: The proposed model improves on existing baselines in tasks requiring spatial reasoning and grounding reasoning.
DUET: Joint Exploration of User–Item Profiles in Recommendation System (2026.findings-acl)

Copied to clipboard

Challenge: Existing LLMs are opaque and difficult to interpret, resulting in limited interpretability.
Approach: They propose an interaction-aware profile generator that jointly produces user and item profiles conditioned on both user history and item evidence.
Outcome: The proposed model outperforms baselines on three real-world datasets.
The Right Time Matters: Data Arrangement Affects Zero-Shot Generalization in Instruction Tuning (2025.findings-acl)

Copied to clipboard

Challenge: Existing work on instruction tuning has focused on task level, without considering that tasks are artificially defined and, to LLMs, merely consist of tokens and representations.
Approach: They propose a training data arrangement framework that allows for continual learning and loss reduction.
Outcome: The proposed framework promotes continual learning and loss reduction on unseen tasks.
Amanda: Adaptively Modality-Balanced Domain Adaptation for Multimodal Emotion Recognition (2024.findings-acl)

Copied to clipboard

Challenge: Emotion recognition is a multimodal learning method that can be used for data scarcity.
Approach: They propose to use Adaptively modality-balanced domain adaptation to balance the alignment of different modalities for multimodal emotion recognition.
Outcome: The proposed model outperforms competing models on common datasets on multimodal emotion recognition.
Third-Person Appraisal Agent: Simulating Human Emotional Reasoning in Text with Large Language Models (2025.findings-emnlp)

Copied to clipboard

Challenge: Emotional reasoning is essential for improving human-AI interactions, especially in mental health support and empathetic systems.
Approach: They propose a third-person appraisal agent that simulates human-like emotional reasoning through three phases: Primary Appraisal, Secondary Appraisals, and Reappraisal.
Outcome: The proposed model outperforms baseline LLMs in various emotional reasoning tasks, demonstrating superior generalization and interpretability.
User Feedback Alignment for LLM-powered Exploration in Large-scale Recommendation Systems (2025.acl-industry)

Copied to clipboard

Challenge: Large Language Models (LLMs) can be used to broaden user experiences beyond established preferences and reinforce feedback loops.
Approach: They propose a hierarchical approach that combines hierarchic planning with LLM inference-time scaling to improve recommendation relevancy without compromising novelty.
Outcome: The proposed approach shows significant gains in both user satisfaction and exploration diversity.
GRADA: Graph-based Reranking against Adversarial Documents Attack (2025.emnlp-main)

Copied to clipboard

Challenge: Retrieval Augmented Generation (RAG) frameworks are susceptible to adversarial attacks that manipulate the retrieval process by introducing documents that are adversarially similar to the query.
Approach: They propose a framework that integrates external retrieval modules into RAG frameworks to improve the factual accuracy of large language models.
Outcome: The proposed framework reduces adversarial attacks by 80% while maintaining minimal loss in accuracy.
DetectiveNN: Imitating Human Emotional Reasoning with a Recall-Detect-Predict Framework for Emotion Recognition in Conversations (2024.findings-emnlp)

Copied to clipboard

Challenge: Existing methods for Emotion Recognition in conversations are insufficient in understanding the rich historical emotional context.
Approach: They propose a novel model that utilizes a "recall-detect-predict" framework to imitate human emotional reasoning by 'recalling' past interactions of a speaker to collect emotional cues.
Outcome: The proposed model outperforms existing methods on three benchmark datasets and significantly outperformed existing methods.
CVAE-based Re-anchoring for Implicit Discourse Relation Classification (2021.findings-emnlp)

Copied to clipboard

Challenge: Existing studies show that training implicit discourse relation classifiers suffers from data sparsity.
Approach: They propose a re-anchoring strategy to reduce the risk of erroneous sampling . they use Conditional VAE to estimate the risk and migrate the anchor to reduce it .
Outcome: The proposed method improves the baseline classifier performance on PDTB v2.0 .
Parrot: A Training Pipeline Enhances Both Program CoT and Natural Language CoT for Reasoning (2025.emnlp-main)

Copied to clipboard

Challenge: Existing work focuses on enabling models to generate natural language chain-of-thought rationales or leverage executable and verifiable code, such as Python.
Approach: They propose a novel training pipeline that integrates sequential P-CoT and N-Co T generation and a subtask hybrid training strategy to facilitate natural language transferability.
Outcome: The proposed training pipeline improves both N-CoT and P-Co T performance over the RL baseline.
Infinite Babble: Inflating 3D Vision-Language Model Inference Overhead via Adversarial Geometric Perturbation (2026.findings-acl)

Copied to clipboard

Challenge: 3D Vision-Language Models (VLMs) are critical cognitive backbone for spatial intelligence, but their reliance on autoregressive decoding introduces a fundamental vulnerability regarding inference efficiency.
Approach: They propose a framework that triggers computational and economic exhaustion in 3D-VLMs by injecting imperceptible noise that forces the model into a state of pathological verbosity.
Outcome: The proposed framework amplifies output length and energy consumption by up to 6.45, demonstrating a potent capability to drain system resources.
CoLLiE: Collaborative Training of Large Language Models in an Efficient Way (2023.emnlp-demo)

Copied to clipboard

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.
Preference Heads in Large Language Models: A Mechanistic Framework for Interpretable Personalization (2026.acl-long)

Copied to clipboard

Challenge: Large Language Models exhibit strong implicit personalization ability, but most approaches treat this behavior as a black box.
Approach: They propose a mechanistic interpretation perspective and propose 'sparse' set of Preference Heads . they compute a Preference Contribution Score for each attention head and compare their predictions .
Outcome: The proposed framework computes a Preference Contribution Score (PCS) for each attention head and measures its causal impact on user aligned outputs.
Why Steering Works: Toward a Unified View of Language Model Parameter Dynamics (2026.acl-long)

Copied to clipboard

Challenge: Methods for controlling large language models (LLMs) are often studied in isolation, obscuring connections and making comparison difficult.
Approach: They propose a preference-utility analysis that separates control effects into preference and utility, and measures both on a shared log-odds scale using polarity-paired contrastive examples.
Outcome: The proposed approach improves preference while preserving utility.
Few-shot Joint Multimodal Aspect-Sentiment Analysis Based on Generative Multimodal Prompt (2023.findings-acl)

Copied to clipboard

Challenge: Existing studies require massive labeled data to train models for multimodal data analysis.
Approach: They propose a novel multimodal prompt model that captures specific aspect terms in a few-shot scenario.
Outcome: The proposed model outperforms baselines on two MABSA-related tasks on a few-shot dataset.

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