Team Members

Each of our researchers are experts in emerging technologies such as machine learning algorithms, forecasting models, intelligent image and video search, automated scene analysis, voice biometrics, recommender systems, personalization, and deep metadata. Combined they hold over 100 US patents across their areas of specialty, and have contributed to and presented more than 50 research projects.

We are a distributed team, with locations across the United States such as Washington, DC,  Philadelphia, Chicago, Denver and Silicon Valley,

Group Lead

Jan Neumann, Ph.D., Senior Director

Applied Machine Learning & Data Science

Bernard Burg, Ph.D., Senior Manager

Oliver Jojic, Distinguished Researcher

Fan Liu, Ph.D., Lead Data Scientist

Michael Kreisel, Ph.D., Lead Researcher

Ryan March, Senior Research Engineer

Nicholas Pinckernell, Senior Principal Research Engineer

Sunil Srinivasa, Ph.D., Lead Researcher

Abel Villca Rocque, Ph.D, Principal Researcher

Natural Language Processing & Content Discovery

Lead: Hassan Sayyadi, Ph.D., Director

Nima Bina, Lead Engineer

Anupam Guha, Ph.D., Senior Researcher

Sahil Jambhekar, Senior Research Engineer

Craig Murray, Manager Engineering

Raul Guerra Paredes, Lead Engineer

Robert Rubinoff, Ph.D., Senior Principal Researcher

Vikrant Sagar, Lead Engineer NLP

Shahin Sefati, Ph.D., Manager, Content Discovery Research

Yihui Tang, Ph.D., Lead Research Engineer

Ferhan Ture, Ph.D., Manager, NLP Research

Richard Walsh, Lead Researcher

Andy Wetta, Ph.D., Lead Researcher

Media Analytics

Lead: Faisal Ishtiaq, Ph.D., Director

Tony Braskich, Senior Research Engineer

Mahmudul Hasan, Ph.D., Senior Researcher

Richard Li, Ph.D., Senior Researcher

Isselmou Maloum, Ph.D., Senior Researcher

Ehsan Younessian, Ph.D., Lead Researcher

Digital Home (Video, Sensor & WiFi Analytics)

Lead: Hongcheng Wang, Ph.D., Senior Manager

Tianwen Chen, Ph.D., Lead Data Scientist

Navdeep Jain, Senior Research Engineer

Abhijeet Mulye,Lead Researcher

Toufiq Parag, Ph.D., Principal Researcher

Vamsi Potluru, Ph.D., Lead Researcher

Scott Rome, Ph.D., Lead Researcher

Donald Tolley, Lead Computer Vision Engineer

Yonatan Vaizman, Ph.D., Senior Researcher

Research Interns 2018

Zhiyuan Cao, UCLA

Marshall Grant, Georgia State University

Shriya Gupta, University of Maryland, College Park

Aya Ismail, University of Maryland, College Park

Ashish Ranjan – UMass, Amherst

Parsa Saadatpanah, University of Maryland, College Park

Ralph Tang, University of Waterloo

Hong Wei, University of Maryland, College Park

Yuanwei Wu, University of Kansas

Zhe Wu, University of Maryland, College Park

Research Interns 2017

Sardar Hamidian, George Washington, University

Mohamed Ibrahim, Rutgers University

Upal Mahbub, University of Maryland, College Park

Md Iftekhar Tanveer, University of Rochester

Jinfeng Rao, University of Maryland, College Park

Parsa Saadatpanah, University of Maryland, College Park

Weiwei Yang, University of Maryland, College Park

Ruichi Yu, University of Maryland, College Park

Comcast Applied Artificial Intelligence Research

The Comcast Applied AI Research Team invents the technological foundations for the Xfinity experiences of the future. Besides long-term research efforts in NLP, computer vision, and machine learning (with a focus on deep learning), we support the Comcast Technology, Products and Experience organization through innovations and technical expertise in these product domains:

“Let our customers interact with all our products on their terms”

  • Voice: Communicate with your TV, mobile device and home using natural language.
  • Virtual Assistants and Conversational Dialog: Let customers interact with us on their terms however they prefer.

“Help our customers find and navigate the content they love”

  • Scene-Level Metadata: Understand content at a deeper level to label and categorize the moments and segments of interest with computer vision and machine learning.
  • Content Discovery: Improve the quality of search and recommendations using personalization, predictive analytics, and related advanced machine learning techniques.

“Make our customer’s home secure and convenient and provide the best internet experience”

  • Smart Home Video Analytics & IoT:  Build intelligent product thats secure your home and anticipate your needs using machine learning.
  • Smart WiFI & Internet: Optimize the performance of your in-home and out-home connectivity using machine learning

“Reinvent customer care using AI and machine learning”

  • Proactive Care: Improve customer service experiences or automatically detect and repair outages before they happen with advanced machine learning models.
  • Intelligent Problem Solving: Find the best solution to customer and application questions and issues via smart decision services (bandit and reinforcement learning)

Selected Publications & Talks

The Comcast Applied AI Researchers are experts in their fields of study and are sought after to present at academic and industry conferences:

Deep Learning for Smart Home Monitoring. Hongcheng Wang, GPU Technical Conference (GTC), Oct. 23-24, Washington, D.C., 2018 [Link]

How Comcast uses AI to reinvent the customer experience. Jan Neumann. CableLabs Summer Conference, Keystone, CO, August 7th, 2018

Multi-Perspective Relevance Matching with Hierarchical ConvNets for Social Media Search.  Jinfeng Rao, Wei Yang, Yuhao Zhang, Ferhan Ture, and Jimmy Lin. arXiv (Submitted 5/21/2018) [pdf]

Multi-Task Learning with Neural Networks for Voice Query Understanding on an Entertainment Platform. Jinfeng Rao, Ferhan Ture, and Jimmy Lin, In Proc. of International Conference on Knowledge Discovery & Data Mining (KDD 2018). [pdf]

How Voice helps Comcast to reinvent the customer experience. Jeanine Heck & Jan Neumann, VOICE Summit, Newark, NJ, July 2018,  [Video]

What Do Viewers Say to Their TVs? An Analysis of Voice Queries to Entertainment Systems.  Jinfeng Rao, Ferhan Ture, and Jimmy Lin, In Proc. of ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR 2018). [pdf]

Representing Videos based on Scene Layouts for Recognizing Agent-in-Place Actions. Ruichi Yu, Hongcheng Wang, Ang Li, Jingxiao Zheng, Vlad I. Morariu and Larry Davis, arXiv:1804.01429 [cs.CV] [PDF]

Architecting a Smart Home Monitoring System with Millions of Cameras. Hongcheng Wang, Embedded Vision Summit, Santa Clara, CA, May 22, 2018 [Link]

How Comcast uses AI to reinvent the customer experience. Jan Neumann & Dominique Izbicki. AI Conference NYC, May 1st, 2018. [Link]

How Comcast Uses Deep Learning to Build Intelligent Products and Applications. Jan Neumann. Nvidia GPU Technology Conference, San Jose, CA, March 2018 [Link]

ReMotENet: Efficient Relevant Motion Event Detection for Large-scale Home Surveillance Videos. Ruichi Yu, Hongcheng Wang and Larry Davis, IEEE WACV, Stateline, NV, arXiv:1801.02031, 2018 [PDF]

Talking to Your TV: Context-Aware Voice Search with Hierarchical Recurrent Neural Networks. Jinfeng Rao, Ferhan Ture, Hua He, Oliver Jojic, and Jimmy Lin, In Proc. of International Conference on Information and Knowledge Management (CIKM 2017) 

No Need to Pay Attention: Simple Recurrent Neural Networks Work! (for Answering “Simple” Questions), Ferhan Ture and Oliver Jojic, In Proc. of Empirical Methods in NLP (EMNLP 2017)

Integrating Lexical and Temporal Signals in Neural Ranking Models for Searching Social Media Streams. Jinfeng Rao, Hua He, Haotian Zhang, Ferhan Ture, Royal Sequiera, Salman Mohammed, and Jimmy Lin, To appear in SIGIR 2017 Workshop on Neural Information Retrieval (Neu-IR’17) [pdf]

Mining Temporal Statistics of Query Terms for Searching Social Media Posts. Jinfeng Rao, Ferhan Ture, Xing Niu and Jimmy Lin, To appear in International Conference on the Theory of Information Retrieval (ICTIR 2017)

How AI Powers the Comcast X1 Voice Interface, Jan Neumann, Ferhan Ture & Oliver Jojic, O’Reilly AI Conference, June 2017, New York NY [Conference link]

Data Science & Machine Learning to Improve the Customer Experience, Jan Neumann, Business Analytics Innovation Summit,May 2017, Chicago,IL [Conference Schedule]

How GPUs Power Comcast’s X1 Voice Remote and Smart Video Analytics, Jan Neumann, GPU Technology Conference, May 2017, San Jose, CA [Conference Link]

How AI Powers the X1 Entertainment System. Jan Neumann, AI Summit NYC, Dec 2016, New York, NY [Conference web site]

Learning to Translate for Multilingual Question Answering. Ferhan Ture and Elizabeth Boshee, 2016 Conference on Empirical Methods in Natural Language Processing, Austin, TX [PDF]

3rd Workshop on Recommender Systems for TV and Online Video (RecSysTV 2016) organized by Jan Neumann, John Hannon, and Hassan Sayyadi  in conjunction with ACM RecSys 2016, Boston, MA

Ask Your TV: Real-Time Question Answering with Recurrent Neural Networks. Ferhan Ture and Oliver Jojic, SIGIR 2016, Pisa, Italy [ACM DL] [Arxiv pdf]

Learning Temporal Regularity in Video Sequences. Mahmudul Hasan, Jonghyun Choi, Jan Neumann, Amit K. Roy-Chowdhury, Larry S. Davis, CVPR 2016. [PDF] [Supplementary] [Project Page] [Code] [Video1] [Video2] [Poster]
arXiv:1604.04574 [PDF-Preprint]

How to Fake Multiply with a Gaussian Matrix. Michael Kapralov, Vamsi K. Potluru and David P. Woodruff, ICML 2016. (paper, code, arxiv)

Structured TV Shows — “You have been Chopped”. Ferhan Ture, Jonghyun Choi, Hongcheng Wang and Vamsi K. Potluru, ICML workshop 2016. (pdf)

Data Science and Machine Learning to Improve the Customer Experience. Jan Neumann, Business Analytics Summit – Innovation Enterprise, Chicago, IL, May 2016

How Automatic Content Analytics enables the TV Experiences of the Future. Jan Neumann, INTX 2016, Boston, MA

How Comcast Uses Data Science and ML to Improve the Customer Experience. Jan Neumann, Global Big Data Conference March 2016, Santa Clara, US

Knowledge Transfer with Interactive Learning of Semantic Relationships. Jonghyun Choi, Sung Ju Hwang, Leonid Sigal and Larry S. Davis,
AAAI Conference on Artificial Intelligence (AAAI) 2016 (PDF)

How Spark is working out at Comcast Scale. Sridhar Alla and Jan Neumann, Strata Hadop NYC 2015

Recommendations for Live TV. Jan Neumann and Hassan Sayyadi, RecSys 2015, Vienna Austria

2nd Workshop on Recommender Systems for TV and Online Video (RecSysTV 2015) organized by Jan Neumann, Hassan Sayyadi, John Hannon, Roberto Turrin, and Danny Bickson in conjunction with ACM RecSys 2015, Vienna, Austria

Data Science for Customer Service at Comcast. Jan Neumann, Dato Data Science Summit, San Francisco 2015 (Video)

Comcast – Real-time Recommendations with Spark. Jan Neumann, DC Spark Interactive Meetup, Washington, DC, May 2015 (Video)

Real-time Recommendations using Apache Spark. Jan Neumann &  Sirdhar Alla, Spark Summit East 2015 (Video)

1st Workshop on Recommender Systems for TV and Online Video (RecSysTV 2014) organized by Jan Neumann, Hassan Sayyadi, John Hannon, Roberto Turrin, and Danny Bickson in conjunction with ACM RecSys 2014, Foster City, USA

A probabilistic definition of item similarity. Oliver Jojic, Manu Shukla and Niranjan Bhosarekar. ACM Recsys 2011, Chicago, IL, USA. [pdf]