Papers by Zhe Gan
Adversarial Domain Adaptation for Machine Reading Comprehension (D19-1)
Copied to clipboard
| Challenge: | Existing models for machine reading comprehension rely on large amounts of human-annotated in-domain data. |
| Approach: | They propose an unsupervised domain adaptation framework for Machine Reading Comprehension where the source domain has a large amount of labeled data, while only unlabeled passages are available in the target domain. |
| Outcome: | The proposed framework can be generalizable to different MRC models and datasets and can be extended to semi-supervised learning. |
TrInk: Ink Generation with Transformer Network (2025.emnlp-main)
Copied to clipboard
Zezhong Jin, Shubhang Desai, Xu Chen, Biyi Fang, Zhuoyi Huang, Zhe Li, Chong-Xin Gan, Xiao Tu, Man-Wai Mak, Yan Lu, Shujie Liu
| Challenge: | Existing methods for handwriting generation capture global dependencies and can generate high-quality handwritten samples. |
| Approach: | They propose a Transformer-based model for ink generation, TrInk, which captures global dependencies. |
| Outcome: | The proposed model reduces character error rate and word error rate by 35.56% on the IAM-OnDB dataset compared to previous models. |
EarlyBERT: Efficient BERT Training via Early-bird Lottery Tickets (2021.acl-long)
Copied to clipboard
| Challenge: | Large-scale pre-trained language models require enormous computational resources and long training time. |
| Approach: | They propose an algorithm to reduce inference time and train large NLP models by slimming the self-attention and fully-connected sub-layers inside a transformer. |
| Outcome: | The proposed algorithm achieves comparable performance to standard BERT with 35 45% less training time. |
Improve Vision Language Model Chain-of-thought Reasoning (2025.acl-long)
Copied to clipboard
Ruohong Zhang, Bowen Zhang, Yanghao Li, Haotian Zhang, Zhiqing Sun, Zhe Gan, Yinfei Yang, Ruoming Pang, Yiming Yang
| Challenge: | Current training recipes often rely on datasets dominated by short annotations with limited rationales, hindering the models' ability to generalize to tasks requiring comprehensive reasoning. |
| Approach: | They propose a two-stage post-training strategy that augments short answers with CoT reasoning generated by GPT-4o, enhancing the VLM's CoT capabilities through fine-tuning. |
| Outcome: | The proposed strategy enhances the model's CoT capabilities through fine-tuning and reinforcement learning. |
Cluster-Former: Clustering-based Sparse Transformer for Question Answering (2021.findings-acl)
Copied to clipboard
| Challenge: | Existing models for encoding long sequences in deep learning suffer from high latency and memory demands. |
| Approach: | They propose a clustering-based sparse Transformer framework to perform attention across chunked sequences. |
| Outcome: | The proposed framework achieves state-of-the-art on several major QA benchmarks. |
Distilling Knowledge Learned in BERT for Text Generation (2020.acl-main)
Copied to clipboard
| Challenge: | Large-scale pre-trained language models such as BERT have revolutionized the state of the art in many language understanding tasks. |
| Approach: | They propose a conditional masked language modeling approach to fine tune BERT on target generation tasks by imposing global sequence-level supervision on conventional Seq2Seq models. |
| Outcome: | The proposed model outperforms strong Transformer baselines on multiple language generation tasks such as machine translation and text summarization. |
Contrastive Distillation on Intermediate Representations for Language Model Compression (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods to compress language models use a simple L_2 loss to distill knowledge in the intermediate representations of a large BERT model to a smaller one. |
| Approach: | They propose a method that uses knowledge distillation to distill knowledge through intermediate layers of the teacher via a contrastive objective. |
| Outcome: | The proposed method outperforms state-of-the-art methods on the GLUE benchmark. |
Topic-Guided Variational Auto-Encoder for Text Generation (N19-1)
Copied to clipboard
Wenlin Wang, Zhe Gan, Hongteng Xu, Ruiyi Zhang, Guoyin Wang, Dinghan Shen, Changyou Chen, Lawrence Carin
| Challenge: | Experimental results show that our model outperforms its competitors on both unconditional and conditional text generation. |
| Approach: | They propose a topic-guided variational auto-encoder model for text generation that specifies a Gaussian mixture model and a neural topic module to generate sentences under the topic. |
| Outcome: | The proposed model outperforms existing variational auto-encoders on unconditional and conditional text generation, and can generate semantically-meaningful sentences with various topics. |
Discourse-Aware Neural Extractive Text Summarization (2020.acl-main)
Copied to clipboard
| Challenge: | Recent studies have shown that sentence-based extractive models result in redundant or uninformative phrases in the extracted summaries. |
| Approach: | They propose a discourse-aware neural summarization model that extracts sub-sentential discourse units as candidates for extractive selection on a finer granularity. |
| Outcome: | Experiments show that the proposed model outperforms state-of-the-art models on popular summarization benchmarks. |
An Empirical Study of Multimodal Model Merging (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Existing studies have shown that model merging can generate a multi-task solution without synchronous training. |
| Approach: | They propose to merge vision, language, and cross-modal transformers of a modality-specific architecture to create a parameter-efficient architecture. |
| Outcome: | The proposed model merging outperforms naive models on various tasks with improvements of 3% on VQA, 7% on COCO retrieval, 25% on NLVR2, 14% on Flickr30k and 3% ADE20k. |
Cross-Thought for Sentence Encoder Pre-training (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing models to pretrain sentence encoders with large unlabeled corpus are lacking in linguistic information retrieval. |
| Approach: | They propose a novel approach to pre-training sequence encoder using transformers . they propose to train a Transformer-based sequence encoded over a large set of short sequences based on a set of masked words . |
| Outcome: | The proposed approach outperforms state-of-the-art encoders on hotpotQA by improving intermediate information retrieval performance. |
Pre-trained Language Models Do Not Help Auto-regressive Text-to-Image Generation (2024.emnlp-main)
Copied to clipboard
| Challenge: | Recent advances in image tokenizers have enabled text-to-image generation using auto-regressive methods, but these methods lack pre-trained language models for text-based models. |
| Approach: | They adapt a pre-trained language model for auto-regressive text-to-image generation and show that pre-train language models offer limited help. |
| Outcome: | The proposed model is compared with a pre-trained language model and shows that it is no more effective than random initialized models. |
Multi-step Reasoning via Recurrent Dual Attention for Visual Dialog (P19-1)
Copied to clipboard
| Challenge: | Existing models for visual dialog infer the answer through multiple reasoning steps. |
| Approach: | They propose a model for visual dialog that uses multi-step reasoning to answer questions about an image. |
| Outcome: | The proposed model achieves a new state-of-the-art of 64.47% on the VisDial v1.0 dataset . |
Hierarchical Graph Network for Multi-hop Question Answering (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing multi-hop question answering models focus on multi-level reasoning across multiple documents or paragraphs. |
| Approach: | They propose a hierarchical graph network that aggregates clues from scattered texts . they use a set of contextual encoders to initialize nodes on different levels of granularity . |
| Outcome: | The proposed model outperforms existing multi-hop QA approaches on the HotpotQA benchmark. |
POINTER: Constrained Progressive Text Generation via Insertion-based Generative Pre-training (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing pre-trained language models cannot be directly employed to generate text under specified lexical constraints. |
| Approach: | They propose a method for insertion-based text generation that inserts tokens between existing tokens in a parallel manner. |
| Outcome: | The proposed method is intuitive and interpretable on Wikipedia and Yelp datasets. |
APo-VAE: Text Generation in Hyperbolic Space (2021.naacl-main)
Copied to clipboard
| Challenge: | Existing models that learn embeddings only in Euclidean vector space do not account for such structural property of language. |
| Approach: | They propose a Poincare Variational Autoencoder to capture latent hierarchies in hyperbolic space . they propose enabling adversarial learning procedures to empower robust model training . |
| Outcome: | The proposed model outperforms existing models in a hyperbolic latent space . it captures latent language hierarchies in hyperbolical space and is robust to training . |
Domain Adaptive Text Style Transfer (D19-1)
Copied to clipboard
| Challenge: | Text style transfer without parallel data is a promising method for learning, but in the scenario where less data is available, it may yield poor performance. |
| Approach: | They propose to leverage available data to learn domain-adaptive text style transfer models . they evaluate two style transfer tasks where only limited non-parallel data is available . |
| Outcome: | The proposed models learn from the source domain to: (i) distinguish stylized information and generic content information; (ii) maximally preserve content information and (iv) adaptively transfer the styles in a domain-aware manner. |
Patient Knowledge Distillation for BERT Model Compression (D19-1)
Copied to clipboard
| Challenge: | Pre-trained language models such as BERT have proven to be highly effective for natural language processing tasks, but the high demand for computing resources hinders their application in practice. |
| Approach: | They propose to compress an original large model (teacher) into an equally-effective lightweight shallow network (student) Empirically, this translates into improved results on multiple NLP tasks with a significant gain in training efficiency, without sacrificing model accuracy. |
| Outcome: | The proposed model reduces the computational cost of training models using the teacher model into a lightweight shallow network. |
TIGEr: Text-to-Image Grounding for Image Caption Evaluation (D19-1)
Copied to clipboard
Ming Jiang, Qiuyuan Huang, Lei Zhang, Xin Wang, Pengchuan Zhang, Zhe Gan, Jana Diesner, Jianfeng Gao
| Challenge: | Existing metrics based on text-level comparisons fail to assess the quality of captions produced by machines. |
| Approach: | They propose to use a machine-learned text-image grounding model to measure the accuracy of machine-generated captions and their correlation with human judgments. |
| Outcome: | The proposed metric has higher consistency with human judgments and is more accurate than existing metrics. |
Contextual Text Style Transfer (2020.findings-emnlp)
Copied to clipboard
| Challenge: | Existing methods for text style transfer are limited by the lack of parallel data. |
| Approach: | They propose a task to translate a sentence into a desired style with its surrounding context taken into account. |
| Outcome: | The proposed model outperforms state-of-the-art methods across style accuracy, content preservation and contextual consistency metrics. |
Multi-Fact Correction in Abstractive Text Summarization (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing abstractive summarization systems generate incorrect facts with respect to the source text. |
| Approach: | They propose a suite of two factual correction models that leverages question-answering knowledge to make corrections in system-generated summaries via span selection. |
| Outcome: | The proposed model improves factuality of news summarization without sacrificing summary quality. |
Improving Adversarial Text Generation by Modeling the Distant Future (2020.acl-main)
Copied to clipboard
Ruiyi Zhang, Changyou Chen, Zhe Gan, Wenlin Wang, Dinghan Shen, Guoyin Wang, Zheng Wen, Lawrence Carin
| Challenge: | Recent work has shown excellent performance on text generation tasks by combining reinforcement learning (RL) and generative models. |
| Approach: | They propose a model-based imitation-learning approach to improve text generation performance by focusing on a long horizon. |
| Outcome: | The proposed model improves on a number of text-generation tasks and provides intermediate rewards for generator optimization. |
HERO: Hierarchical Encoder for Video+Language Omni-representation Pre-training (2020.emnlp-main)
Copied to clipboard
| Challenge: | HERO is a framework for large-scale video+language omni-representation learning. |
| Approach: | They propose a framework for large-scale video+language omni-representation learning that encodes multimodal inputs in a hierarchical structure and uses Masked Language Modeling and Masked Frame Modeling to train models. |
| Outcome: | The proposed framework achieves state-of-the-art on multiple benchmarks over text-based video/video-moment retrieval, video question answering (QA), Video-and-language Inference and video Captioning tasks across different domains. |