I am a researcher and research manager at Vision and Language Lab, Adobe Research. I develop machine learning algorithms for Adobe digital media and digital marketing products. I am interested in applying AI techniques to intelligent applications such as search & recommendation, personal assistant agents, intelligent tutoring, spoken dialogue systems, multimodal interfaces, conversational video games, and AI for healthcare. My work on dialog won the best paper award at SigDial 2018 and the winner of the 4th Dialog State Tracking Challenge 2016. My recent work on AI for healthcare won the 2nd ACM SIGAI Industry Award as one of three Team Leads (presented at IJCAI 2021).
I joined Adobe Research in 2012. Before that, I was at North Side Inc. in Montreal, where I led the language team to develop dialog management technologies for the Bot Colony game (~23 million USD investment). I conducted my Ph.D. research at the Human Media Interaction Group, University of Twente, the Netherlands. I worked as an engineering research associate at Stanford University in the Center for the Study of Language and Information. I started my research career as a research assistant at the Artificial Intelligence Laboratory, Swiss Federal Institute of Technology, in Lausanne, Switzerland.
Program committee member (2019-present): ACL, EMNLP, NAACL, AAAI, CVPR, ICCV
Journal reviewer: TACL 2021-2023, NLE (2021 - present)
News:
2024
Mar 13: 1 paper accepted at NAACL
Feb 26: 1 pager accepted at CVPR
Jan 25: A patent from my team appeared in the Daily Upside.
Jan 18: 1 paper accepted at EACL
Jan 16: 1 paper accepted at ICLR for oral presentation (1.2% acceptance rate over 7262 submissions)
2023
2 papers accepted at EMNLP
3 papers accepted at ICCV
1 paper accepted at Interspeech
2 papers accepted at ACL
2 papers accepted at CVPR (acceptance rate: 2359/9155 = 25.8%)
1 paper accepted at EACL (acceptance rate: 281/1166 = 24.1%)
1 paper accepted at ICASSP
Tech lead and primary contributor for Adobe Firefly support text prompts in over 100 languages and currently being used in Firefly, Photoshop, Express, and Illustrator. See here.
2022
1 paper accepted at AACL-IJNLP (oral, acceptance rate: 63/554 = 11.4%)
1 paper accepted at ACCV (oral, acceptance rate: 41/836 = 4.9%)
4 papers accepted at COLING (acceptance rate: 634/2026 = 31.3%)
1 paper accepted at Interspeech
2 papers accepted at NAACL
1 paper accepted at ICASSP
Key contributor for Behance LiveStream Search released in Q2 2022.
Primary contributor for keyphrase extraction service in Blink Q2 2022
2021
1 paper accepted at AAAI 2022
2 papers accepted at EMNLP
3 papers accepted at ACL
1 paper accepted at CVPR
1 long paper accepted at NAACL (acceptance rate: 350/1245 = 28.1%)
Winner of the 2021 ACM SIGAI Industry Award as Team Lead
Key primary contributor for the Adobe AEM smart tag intelligence service released in Feb 2021
2020
3 papers accepted at EMNLP
1 papers accepted at ACL
1 paper accepted at CVPR
1 paper accepted at COLING
1 paper accepted at ICASSP
Key primary contributor for the cloud-power search suggestions in Acrobat released on May 12, 2020, which serves hundreds of millions of users monthly
2019
1 paper accepted at EMNLP
1 paper accepted at CIKM
1 paper accepted at Interspeech
1 paper accepted at ACL (oral, acceptance rate: 660/2905 = 22.7%)
2018
Invited plenary talk at the 1st Uber Science Symposium, San Francisco, Nov 28, 2018 (details)
Our paper at MULA-ECCV 2018 won the best paper award.
Our paper at SIGDIAL 2018 won the best paper award.
1 paper accepted at COLING
1 paper accepted at Interspeech
2 papers accepted at NAACL
1 paper accepted at CVPR (acceptance rate: 797/3303 = 29.6%)
2017
1 paper accepted at CVPR (acceptance rate: 783/2680 = 29%)
Speech Driven Book Technology code is now available on github with Apache 2.0 license: link
Voice-based photo editing project has received a lot of attractions from the press: Youtube, Forbes, Tech Crunch, Yahoo. Here is the full list of coverage.
2016
Main contributor of the new lifetime value algorithm that has been using in Adobe Target since 2016. Here is the news: Adobe.