Haw-Shiuan Chang

張浩軒

I’m currently in my final year of my PhD at the University of Massachusetts Amherst advised by Professor Andrew McCallum.

My primary research goal is to build accurate semantic models of textual similarity and entailment in order to help knowledge workers efficiently navigate large text collections. To approach this goal, I addressed some fundamental limitations of self-supervised language models and active learning methods; I developed a state-of-the-art review-paper affinity estimation model that is adopted by OpenReview and used by many conferences, including ICLR and NeurIPS; I also developed unsupervised methods for hypernym detection and crowdsourcing methods for educational prerequisite estimation.

Previously, I worked with Yu-Chiang Frank Wang and Kuan-Ta Chen in Academia Sinica, Taiwan. I received BS in EECS Undergraduate Honors Program from National Yang Ming Chiao Tung University (NYCU), Taiwan.

Selected publications

  1. ACL
    Softmax Bottleneck Makes Language Models Unable to Represent Multi-mode Word Distributions
    Haw-Shiuan Chang, and Andrew McCallum
    In Annual Meeting of the Association for Computational Linguistics (ACL) 2022
  2. EACL Oral
    Changing the Mind of Transformers for Topically-Controllable Language Generation
    Haw-Shiuan Chang, Jiaming Yuan, Mohit Iyyer, and Andrew McCallum
    In Conference of the European Chapter of the Association for Computational Linguistics (EACL) (Oral) 2021
  3. ML
    Using Error Decay Prediction to Overcome Practical Issues of Deep Active Learning for Named Entity Recognition
    Haw-Shiuan Chang, Shankar Vembu, Sunil Mohan, Rheeya Uppaal, and Andrew McCallum
    Machine Learning 2020
  4. NAACL
    Distributional Inclusion Vector Embedding for Unsupervised Hypernymy Detection
    Haw-Shiuan Chang, ZiYun Wang, Luke Vilnis, and Andrew McCallum
    In Human Language Technology Conference of the North American Chapter of the Association of Computational Linguistics (HLT/NAACL) 2018
  5. NIPS
    Active Bias: Training a More Accurate Neural Network by Emphasizing High Variance Samples
    Haw-Shiuan Chang, Erik G. Learned-Miller, and Andrew McCallum
    In Advances in Neural Information Processing Systems (NIPS) 2017
  6. EDM Short
    Modeling Exercise Relationships in E-Learning: A Unified Approach
    Haw-Shiuan Chang, Hwai-Jung Hsu, and Kuan-Ta Chen
    In International Conference on Educational Data Mining (EDM) 2015