Scholar
Hamid Palangi
Google Scholar ID: B1lAghgAAAAJ
Google and University of Washington
Citations & Impact
Citations
10,832
H-index
28
i10-index
51
Publications
20
Co-authors
58
list available
Publications
Resume (English only)
Academic Achievements
- Paper 'Deep Sentence Representations for Web Search Engines and Information Retrieval' received IEEE Signal Processing Society Best Paper Award 2018
- Paper 'Improving the Reusability of Pre-trained Language Models in Real-world Applications' received Best Paper Award at IEEE IRI 2023
- Led development of Orca series (Orca, Orca 2), student models learning from ChatGPT and GPT-4
- Multiple papers accepted at top conferences including NeurIPS 2023, ACL 2023, ICML 2023, AAAI 2023
- Co-authored 'Sparks of Artificial General Intelligence? Early experiments with GPT-4'
- Proposed a framework for automated measurement of responsible AI harms in generative AI
- Developed CogEval to evaluate cognitive maps and planning in LLMs
- Introduced GRACE: lifelong model editing with discrete key-value adaptors
- Explored synthetic tasks to reduce hallucination in LLMs and transfer to real-world tasks
Research Experience
- Worked on image captioning with Microsoft Cognitive Services
- Led hate speech detection project at Azure; resulting dataset ToxiGen adopted by Llama2, Code Llama, Orca1/2, phi-1.5/2, Gemma, etc., and used in legal and economics forums
- Contributed to document recommendation system in Microsoft Office
- Worked on New Bing and text-image retrieval at Bing
- Developed CELA, an initial NLP system for understanding legal contracts
- Mentor at Microsoft AI School (2017–2019)
- Mentor at Microsoft AI Residency Program (2019–2020)
- Virtual Infrastructure Chair for ACL 2020
Background
- Staff Research Scientist Manager at Google
- Affiliate Associate Professor at the University of Washington
- Advisor for Koidra
- Senior Member of IEEE; served as Area Editor for IEEE Signal Processing Magazine Newsletter
- Research interests include large language models, generative AI, responsible AI, multimodal models, information retrieval, and toxicity detection in legal and economic texts
