Papers by Yijin Chen
Enhancing Cross-Tokenizer Knowledge Distillation with Contextual Dynamical Mapping (2025.findings-acl)
Copied to clipboard
| Challenge: | Knowledge distillation (KD) approaches focus on homogeneous architectures with identical tokenizers, constraining their applicability in cross-architecture scenarios. |
| Approach: | They propose a framework that uses contextual information to enhance sequence alignment precision and dynamically improves vocabulary mapping. |
| Outcome: | The proposed framework shows significant advantages over existing methods for model compression . it can be used across multiple model families and across multiple benchmarks . |
Conditional Bilingual Mutual Information Based Adaptive Training for Neural Machine Translation (2022.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to improve neural machine translation use token-level adaptive training . however, standard models make predictions on condition of previous contexts . |
| Approach: | They propose a target-context-aware metric which can be supplemented by statistical metrics . they propose an adaptive training approach based on token- and sentence-level CBMI . |
| Outcome: | The proposed model outperforms the Transformer baseline and other similar approaches on English-German and Chinese-English tasks. |
Comments as Natural Logic Pivots: Improve Code Generation via Comment Perspective (2024.findings-acl)
Copied to clipboard
| Challenge: | Existing studies decompose complex tasks into intermediate steps by prompting strategies, such as Chain-of-Thought and its variants. |
| Approach: | They propose to use code comments as natural logic pivot between natural language and code language to boost the code generation ability of code LLMs. |
| Outcome: | The proposed method significantly improves the code pass rate on humanEval and MBPP, while the robustness of the logical comment decoding strategy is higher than the Chain-of-thoughts prompting. |
Investigating Multi-Hop Factual Shortcuts in Knowledge Editing of Large Language Models (2024.acl-long)
Copied to clipboard
| Challenge: | Recent work has demonstrated the power of large language models in recalling knowledge and reasoning. |
| Approach: | They propose to erase shortcut neurons to mitigate the associated risks . 20% of the failures are attributed to shortcuts, they find . |
| Outcome: | The proposed approach reduces failures in multi-hop knowledge editing caused by shortcuts by 20% . |
SED-SFT: Selectively Encouraging Diversity in Supervised Fine-Tuning (2026.acl-short)
Copied to clipboard
| Challenge: | Existing studies have proposed a new approach to optimize for SFT followed by RL . existing studies have suggested a method to optimize SFT for large language models . |
| Approach: | They propose a framework that encourages diversity based on token exploration space. |
| Outcome: | Experiments show that SED-SFT significantly improves generation diversity with a negligible computational overhead increase over CE loss. |
LCS: A Language Converter Strategy for Zero-Shot Neural Machine Translation (2024.findings-acl)
Copied to clipboard
| Challenge: | Existing LT strategies cannot indicate the desired target language on zero-shot translation, i.e., the off-target issue. |
| Approach: | They propose a language converter strategy that embeds the target language into the top encoder layers to mitigate confusion in the encoder and ensures stable language indication for the decoder. |
| Outcome: | The proposed language converter strategy significantly mitigates off-target issue on multiUN, TED, and OPUS-100 datasets. |
Confidence-Aware Scheduled Sampling for Neural Machine Translation (2021.findings-acl)
Copied to clipboard
| Challenge: | Neural machine translation models are trained to maximize the likelihood of the next token given previous golden tokens as inputs, but at the inference stage, golden token is unavailable. |
| Approach: | They propose a scheduled sampling method that randomly replaces groundtruth tokens with predicted ones during training, ignoring real-time model competence. |
| Outcome: | The proposed method outperforms the Transformer and vanilla scheduled sampling on large-scale translations. |
Outdated Issue Aware Decoding for Factual Knowledge Editing (2024.findings-acl)
Copied to clipboard
| Challenge: | Existing knowledge editing methods retain outdated responses for reasoning questions . naively retraining LLMs can be computationally intensive and can lead to catastrophic forgetting . |
| Approach: | They propose a simple yet effective decoding strategy to enhance edited models on reasoning questions. |
| Outcome: | The proposed method outDates ISsue aware deCOding (DISCO) to improve models on reasoning questions. |
CM-Net: A Novel Collaborative Memory Network for Spoken Language Understanding (D19-1)
Copied to clipboard
| Challenge: | Existing models for slot filling and intent detection fail to fully utilize cooccurrence relations between slots and intents, which restricts their potential performance. |
| Approach: | They propose a novel Collaborative Memory Network (CM-Net) that captures slot-specific and intent-specific features in a collaborative manner. |
| Outcome: | The proposed network outperforms existing models on two benchmarks and a self-collected corpus. |
Scheduled Sampling Based on Decoding Steps for Neural Machine Translation (2021.emnlp-main)
Copied to clipboard
| Challenge: | Neural machine translation models are trained to maximize the likelihood of next token given previous golden tokens as inputs, but at the inference stage, golden token is unavailable. |
| Approach: | They propose to use scheduled sampling to replace ground-truth tokens with predicted tokens to bridge the gap between training and inference. |
| Outcome: | The proposed methods outperform the Transformer baseline and vanilla scheduled sampling on three large-scale WMT tasks. |
GCDT: A Global Context Enhanced Deep Transition Architecture for Sequence Labeling (P19-1)
Copied to clipboard
| Challenge: | Existing systems for sequence labeling are limited by shallow connections between consecutive hidden states and insufficient modeling of global information. |
| Approach: | They propose a global context enhanced deep transition architecture for sequence labeling . they deepen the state transition path at each position in a sentence and assign tokens with global representations . |
| Outcome: | The proposed architecture outperforms the best reported results on two standard sequence labeling tasks. |