Scholar
Weiting Tan
Google Scholar ID: hD8E4gYAAAAJ
Johns Hopkins University, Bytedance Seed
Citations & Impact
Citations
451
H-index
5
i10-index
3
Publications
17
Co-authors
12
list available
Contact
Publications
Resume (English only)
Academic Achievements
- Sep 25, 2025: Process-Supervised Reinforcement Learning for Interactive Multimodal Tool-Use Agents released on arXiv.
- Aug 20, 2025: Seeing is Believing: Emotion-Aware Audio-Visual Language Modeling for Expressive Speech Generation accepted to EMNLP 2025!
- May 20, 2025: STAR and SSR accepted to IWSLT 2025!
- Sep 25, 2024: DiffNorm accepted to NeurIPS 2024!
- Paper SSR: Alignment-Aware Modality Connector for Speech Language Models proposes a method to better fuse modalities by segmenting and compressing speech features.
- Paper DiffNorm: Self-Supervised Normalization for Non-autoregressive Speech-to-speech Translation introduces a diffusion-based normalization strategy to simplify data distributions.
Research Experience
- Recently focused on allowing multi-modal agents to seamlessly reason, utilize tools, and interact with users/envs. Also working on unified multimodal understanding and generation LLMs.
Education
- Currently a PhD Candidate in Computer Science at Johns Hopkins University, advised by Prof. Philipp Koehn. Previously, completed Undergraduate and Master’s degrees in Computer Science at JHU.
Background
- Research interests: machine learning and natural language processing; particularly excited by efficient and scalable representation learning methods for cross-modal applications; also interested in derivatives pricing and hedging strategies.
Miscellany
- Email: wtan12 at jhu.edu
