Papers by Jinwook Park
Structural Optimization Ambiguity and Simplicity Bias in Unsupervised Neural Grammar Induction (2024.findings-acl)
Copied to clipboard
| Challenge: | Unsupervised grammar induction models lack analysis of traditional challenges, especially regarding the training loss. |
| Approach: | They propose a method to reduce the parse pool per sentence for loss evaluation using structural bias from pre-trained parsers. |
| Outcome: | The proposed method significantly improves performance while reducing variance and bias toward overly simplistic parses. |
AMXFP4: Taming Activation Outliers with Asymmetric Microscaling Floating-Point for 4-bit LLM Inference (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing research shows unsatisfactory empirical results for microscaling (MX) floating-point (FP) formats. |
| Approach: | They propose a 4-bit asymmetric FP format that handles activation outliers without calibration. |
| Outcome: | The proposed format outperforms MXFP4 by 3% on VQA and rotation-based methods by 1.6% on CSQA. |
Probability Distribution Collapse: A Critical Bottleneck to Compact Unsupervised Neural Grammar Induction (2025.emnlp-main)
Copied to clipboard
| Challenge: | Existing models face expressiveness bottlenecks, resulting in unnecessarily large yet underperforming grammars. |
| Approach: | They propose a method to reduce the expressiveness bottleneck of unsupervised neural grammar induction by leveraging neural parameterization to estimate prob-ability distributions. |
| Outcome: | The proposed approach significantly improves parsing performance while enabling the use of significantly more compact grammars across a wide range of languages. |
Feature Difference Makes Sense: A medical image captioning model exploiting feature difference and tag information (2020.acl-srw)
Copied to clipboard
| Challenge: | Existing methods for medical image captioning are limited and lack diversity . current methods do not generalize well when applied to unfamiliar images . |
| Approach: | They propose a feature difference and tag information combined long short-term memory model for chest x-ray report generation. |
| Outcome: | The proposed model outperforms existing models in chest x-ray report generation. |
RSCF: Relation-Semantics Consistent Filter for Entity Embedding of Knowledge Graph (2025.acl-long)
Copied to clipboard
| Challenge: | Knowledge graph embeddings suffer from incompleteness, a problem that is often overlooked . a generalized plug-in approach to SFBR disrupts consistency by concentrating embeddables under entity-based regularization . |
| Approach: | They propose a plug-in KGE method that uses relation specific entity transformation to enhance semantic consistency. |
| Outcome: | The proposed method outperforms state-of-the-art methods in knowledge graph embedding tasks . the proposed method is based on a plug-in approach that disrupts consistency . |