|"Holding Machine Learning Systems Accountable via Explanations" |
|Konuşmacı ||: ||Assoc. Prof. Dr. Muhammad Aurangzeb Ahmad |
|Tarih ||: ||25 Şubat 2019 (Pazartesi) |
|Saat ||: ||15:30 |
|Yer ||: ||Elektrik ve Elektronik Fakültesi, Oda: 1302 |Seminer Konusu:
Machine Learning and AI based systems are increasingly being integrated into the socio-technical fabric of society. While the predictive and discriminative power of such systems offer potentially vast benefits to society, there is also a downside where use of machine learning models can to institutionalized discrimination, false diagnosis, loss of privacy etc. Thus, there have been increasing demands to hold Machine Learning systems accountable. In this talk I will survey the field of machine learning explanations within the purview of accountability of machine learning systems. Limitations of post-hoc vs. ante-hoc models, risk vs. interpretability trade-off, evaluating explanatory models etc. will also be covered.
Muhammad Aurangzeb Ahmad is an Affiliate Associate Professor in the Department of Computer Science at University of Washington and the Principal Data Scientist at KenSci Inc, a startup focused on Artificial Intelligence in Healthcare. He is also an official advisor on AI and Machine Learning to the ministry of Science and Technology in Maldives. Muhammad Aurangzeb has written more than 50 research papers on machine learning, artificial intelligence and computational social science including two best paper awards. In the past he was also a Visiting Research Scientist at the Indian Institute of Technology – Kanpur and was a research affiliate in the Center for Cognitive Science at University of Minnesota. He has a PhD and masters in Computer Science from the Department of Computer Science at the University of Minnesota, his undergraduate degree is also in Computer Science from Rochester Institute of Technology. His doctoral thesis was on modeling human behavior in massive online games. Most recently Muhammad Aurangzeb was on multiple panels on AI Governance at the Global Governance summit in Dubai. His research work has been extensively covered in the media in the West. His current research is in interpretable machine learning, application of Artificial Intelligence and Machine Learning in Healthcare, Massive Historical Social Networks and Personality Emulation.