Yi-Lin Tuan

You can call me Pascal.

my_pic.jpg

I received Ph.D. in Computer Science from the University of California Santa Barbara, where I worked with professor William Wang. Before that, I received B.S. in Electrical Engineering from National Taiwan University, where I was fortunatedly advised by professor Lin-shan Lee, Hung-yi Lee, and Yun-Nung Chen.

I was a research intern at Meta GenAI (2023) working on Safe LLM control with Dan Bikel; a research intern at Meta AI (2022) on state-space model for LLM pretraining with Wenhan Xiong; a applied scientist intern at Amazon Alexa (2021) on knowledge deep reasoning for dialogue system with Sajjad Beygi and Maryam Fazel-Zarandi; a research intern at Meta AI (2020) on quality estimation of machine translation with Ahmed El-Kishky.

I’m building OrchatAI (free for sign-up and play).

For any question, feel free to reach out to me at pascal at orchat dot ai

Selected Awards
  • Top Reviewer Recognition at NeurIPS 2024
  • Best Poster Award at ACM Augmented Humans'22
  • Best Paper Award at SoCalNLP 2019
  • National Taiwan University Electrical Engineering 1960 alumni scholarship, 2016
  • National Taiwan University Electrical Engineering 1960 alumni scholarship, 2014
  • Dean's list at National Taiwan Unversity (top 1 out of 169 students in Electrical Engineering), 2014
Projects

View full publication list

Conversational models with relationships, knowledge graph, personality, reasoning, interpretability, and multi-agent interactions (2018-now)
  • Towards Safety and Helpfulness Balanced Responses via Controllable Large Language Models. Yi-Lin Tuan, Xilun Chen, Eric Michael Smith, Louis Martin, Soumya Batra, Asli Celikyilmaz, William Yang Wang, and Daniel M. Bikel. preprint, 2024
  • CausalDialogue: Modeling Utterance-level Causality in Conversations. Yi-Lin Tuan, Alon Albalak, Wenda Xu, Michael Saxon, Connor Pryor, Lise Getoor, and William Yang Wang. In ACL-Findings, 2023
  • Towards Large-Scale Interpretable Knowledge Graph Reasoning for Dialogue Systems. Yi-Lin Tuan, Sajjad Beygi, Maryam Fazel-Zarandi, Qiaozi Gao, Alessandra Cervone, and William Yang Wang. In ACL-Findings, 2022
  • Local explanation of dialogue response generation. Yi-Lin Tuan, Connor Pryor, Wenhu Chen, Lise Getoor, and William Yang Wang. NeurIPS, 2021. With Python Package: LERG. Installed by pip install lerg
  • Knowledge injection into dialogue generation via language models. Yi-Lin Tuan, Wei Wei, and William Yang Wang. preprint, 2020
  • DyKgChat: Benchmarking Dialogue Generation Grounding on Dynamic Knowledge Graphs. Yi-Lin Tuan, Yun-Nung Chen, and Hung-Yi Lee. In EMNLP-IJCNLP, 2019
  • Personalized Dialogue Response Generation Learned from Monologues. Feng-Guang Su*, Aliyah R Hsu*, Yi-Lin Tuan, and Hung-Yi Lee. In Interspeech, 2019
Reinforcement learning and reward models for language modeling (2016-now)
  • Algorithm Development
    • Flexible Attention-Based Multi-Policy Fusion for Efficient Deep Reinforcement Learning. Zih-Yun Chiu*, Yi-Lin Tuan*, William Yang Wang, and Michael C Yip. In NeurIPS, 2023
    • Quality Estimation without Human-labeled Data. Yi-Lin Tuan, Ahmed El-Kishky, Adithya Renduchintala, Vishrav Chaudhary, Francisco Guzmán, and Lucia Specia. In EACL, 2021
    • Proximal policy optimization and its dynamic version for sequence generation. Yi-Lin Tuan*, Jinzhi Zhang*, Yujia Li, and Hung-yi Lee. preprint, 2018
    • Improving conditional sequence generative adversarial networks by stepwise evaluation. Yi-Lin Tuan, and Hung-Yi Lee. IEEE/ACM Transactions on Audio, Speech, and Language Processing (Journal), 2018
  • Application
    • Knowledge injection into dialogue generation via language models. Yi-Lin Tuan, Wei Wei, and William Yang Wang. preprint, 2020
    • Personalized Dialogue Response Generation Learned from Monologues. Feng-Guang Su*, Aliyah R Hsu*, Yi-Lin Tuan, and Hung-Yi Lee. In Interspeech, 2019
Physics-inspired machine learning (2021-now)
  • Dynamic Latent Separation for Deep Learning. Yi-Lin Tuan, Zih-Yun Chiu, and William Yang Wang. preprint, 2024. A revision from prior preprints: "Atomized Deep Learning Models. preprint. 2022" and "Modeling Data as Atoms. preprint. 2023".
  • State-space model for LLM pretraining at Meta.
Deep learning algorithms for Speech, Singing, Vision, and Robots
  • Speech and Singing:
    • Transcribing lyrics from commercial song audio: The first step towards singing content processing. Che-Ping Tsai*, Yi-Lin Tuan*, and Lin-shan Lee. In ICASSP, 2018
    • Real-time pitch tracking 3D game in 2016.
  • Vision:
    • Deep Learning–Based Perceptual Stimulus Encoder for Bionic Vision. Lucas Relic, Bowen Zhang, Yi-Lin Tuan, and Michael Beyeler. In ACM Augmented Humans, 2022
    • Dynamic Latent Separation for Deep Learning. Yi-Lin Tuan, Zih-Yun Chiu, and William Yang Wang. preprint, 2024. A revision from prior preprint: "Atomized Deep Learning Models. preprint. 2022".
  • Robots:
    • Flexible Attention-Based Multi-Policy Fusion for Efficient Deep Reinforcement Learning. Zih-Yun Chiu*, Yi-Lin Tuan*, William Yang Wang, and Michael C Yip. In NeurIPS, 2023
    • Parallelized reverse curriculum generation. Zih-Yun Chiu, Yi-Lin Tuan, Hung-yi Lee, and Li-Chen Fu. preprint, 2021
    • Piano accompany robot with LEGO EV3 in 2017.