She completed her msc from the university of cambridge as a marshall scholar, her phd from mit, and her postdoc at harvard medical school. Our nonparametric bayesian approach combines model knowledge, inferred from expert information and independent exploration, with policy knowledge inferred from expert trajectories. She works on machine learning, computational statistics and healthcare. Finale doshivelez awarded everett mendelsohn excellence. The paper was coauthored by finale doshivelez and been kim. I am an undergraduate researcher with professor finale doshivelezs data to actionable knowledge lab, and have research experience in other areas of ml as well.
Finale doshivelez, been kim submitted on 28 feb 2017 v1, last revised 2 mar 2017 this version, v2 abstract. In the united states we spend more on health care and get worse outcomes than comparable countries, she said. Finale doshivelez and been kim from autonomous cars and adaptive email lters to predictive policing systems, machine learn ing ml systems are increasingly ubiquitous. Windows dev center developer network technet microsoft developer program channel 9. Directions in cognitive is a program focused on the issues facing cognitive computing consortium members and other ai and cognitive computing applications professionals today, as they scope, design, build, and revise production applications for the enterprise. In the 2010s public concerns about racial and other bias in the use of ai for criminal sentencing decisions and findings of creditworthiness may have led to increased demand for transparent artificial intelligence. Citeseerx hidden parameter markov decision processes. Executive director center for technology and civic life. Transparency and explainability berkman klein center. As a result, many academics and organizations are developing tools to help detect bias in their systems. Journal of the american medical informatics association 24.
Try finale free for 60 days get a finale free trial. Her interests lie at the intersection of machine learning, healthcare, and interpretablity. This software is opensource, released under the terms of the gnu general public license version 3, or any later version of the gpl see copying. Francis bach optimal algorithms for smooth and strongly convex distributed optimization in networks. Aug 30, 2018 insights are drawn from finale doshivelezs talk, a roadmap for the rigorous science of interpretability as well as the paper, towards a rigorous science of interpretable machine learning. Finale doshivelez institute for applied computational. Bayesian nonparametrics with finale doshivelez spring 2017 20152016. Ml software in health care, and proposes a lifecylebased framework given the iterative nature of aiml. Finale doshivelez is an assistant professor of computer science at the harvard paulson school of engineering and applied sciences. Finale doshivelez abstract the partially observable markov decision process pomdp framework has proven useful in planning domains that require balancing actions that increase an agents knowledge and actions that increase an agents reward. In a presentation at the second annual women in data science conference, harvard assistant professor of computer science finale doshi velez shared insights on the overlap between health care, data science and artificial intelligence. Click on the first link on a line below to go directly to a page where finale is defined. Proceedings of the 18th international conference on autonomous agents and multiagent systems.
She completed her phd at mit, her postdoc at harvard medical school, and is a. Finale doshivelez is an assistant professor of computer science at harvard university. A busy spring for the harvard data science initiative the. For any questions regarding flight and accommodation cancellations for wecode 2020, please view the following cancellation advice. Finale doshivelez and na li receive air force young. Application of big data to pediatric safety studies, sept. Towards a rigorous science of interpretable machine learning. Few details have emerged, but its reportedly the first fatality. Reinforcement learning for healthcare finale doshi. Inference, design of experiments, and experimentation in an automated loop.
The challenge of regulating clinical decision support software after 21 st century cures. Her core research in machine learning, computational statistics, and data science is inspired byand often applied tothe objective of accelerating scientific progress and practical impact in healthcare and other domains. Her core work focuses on developing novel machine learning techniques to identify what treatments will work. Finale doshivelez, assistant professor of computer science at the harvard john a. It ultimately finds that, at least for now, ai systems can and should be held to a.
She completed her msc from the university of cambridge as a marshall scholar, her phd from mit, and her. Finale doshivelez, assistant professor of computer science seas. Her interests lie at the intersection of machine learning and healthcare. Sonali parbhoo, maurizio zazzi, volker roth, xuefeng peng. Finale doshivelez agent speaker fee booking contact. Finale doshivelez, harvard university women in data science. A generative approach to interpretable feature selection and extraction 2015. Finale doshi velez, mason kortz, ryan budish, christopher bavitz, samuel j. Pradier, weiwei pan, jiayu yao, soumya ghosh, finale doshi velez 2018 while modern neural networks are making remarkable gains in terms of predictive accuracy, characterizing uncertainty over the parameters of these models in a bayesian setting is challenging because. References 1 accelerated gibbs sampling for the indian buffet process finale doshi and zoubin ghahramani, icml 2009.
Massachusetts institute of technology, september 2005 june 2012 thesis. A selfdriving car operated by uber struck and killed a woman sunday in tempe, arizona. Finale doshi velez is from harvard, boston and will be giving a talk. All of my work currently focuses on questions drawn from realworld technical challenges associated with using machine learning to assist in healthcare contexts. Finale doshivelezs research works harvard university. Finale doshi velez is an assistant professor of computer science at harvard university.
Nonparametric exchangeable models such as the indian buffet process. Assistant professor of computer science harvard school of engineering and applied science. Nips 2015 workshop doshivelez 15406 machine learning for. Lecturer on law, harvard law school, and adjunct lecturer on public policy, harvard kennedy school. We found 35 dictionaries with english definitions that include the word finale. I work on techniques to learn interpretable models by incorporating user studies into optimization. Finale doshivelez massachusetts institute of technology. Finale doshi velez is an assistant professor at the harvard paulson school. Technion israel institute of technology, haifa, israel.
If you think that id be a good mentor, please get in touch after reading the faqs below. The ones marked may be different from the article in the profile. Finale doshi velez, an assistant professor of computer science at harvard, whose core. Full schedule mobile friendly multitrack schedule conference program. Andreas geiger probabilistic and deep models for 3d reconstruction. The possibility of explanation finale doshivelez tedxboston. Citeseerx document details isaac councill, lee giles, pradeep teregowda. May 18, 2016 recent years have seen an unprecedented rise in the availability and size of collections of clinical data such as electronic health records. Miryung kim ucla, litao yan, chenlu wang hci for programming language technologies. Explainable and interpretable models in computer vision and machine learning, arxiv in 2017 examples are not enough, learn to criticize. Talk finale doshivelez umd department of computer science. Tested faultdetection and correction software for autonomous rovers.
Towards a rigorous science of interpretable machine. Biography and booking information for finale doshi velez, assistant professor of computer science at harvard. In this class, well first learn about the perspective that jainism, an eastern religion, takes on the nature of happinessboth what it is and how to find it. How humans will lose control of artificial intelligence. I am supported by an nsf grfp fellowship since 2019. Gershman, david obrien, stuart shieber, jim waldo, david weinberger, alexandra wood.
Finale doshivelez education doctor of philosophy, computer science. Towards improving health decisions with reinforcement. Contact all american speakers bureau to inquire about speaking fees and availability, and book the best keynote speaker for your next event. Finale doshivelez is an assistant professor of computer science at the harvard paulson school of engineering and. Finale doshivelez, an assistant professor of computer science at harvard, whose core research includes machine.
Loeb associate professor in computer science at the harvard paulson school of engineering and applied sciences. Steve chong describing current work and future possibilities in plhci. Improving counterfactual reasoning with kernelised dynamic mixing. Finale doshivelez, assistant professor of computer science at harvard university, sits down with host lisa martin at women in data science. Finale doshi velez s 47 research works with 672 citations and 4,325 reads, including. Finale doshi velez, assistant professor of computer science at the harvard john a. These rich data sources present opportunities to apply. With a growing interest in interpretability, there is an increasing need to characterize what exactly we mean by it and how to sensibly compare the interpretability of different approaches.
My most recent summer was spent doing data science under siri search at apple. We consider reinforcement learning in partially observable domains where the agent can query an expert for demonstrations. I joke that we call it nice people in machine learning, says finale doshi velez, a harvard computer science professor and wiml executive board member. Everything is included except the premium garritan sounds. Finale doshi velez, been kim submitted on 28 feb 2017 v1, last revised 2 mar 2017 this version, v2 abstract. Facebook gives people the power to share and makes the world more open and connected. Finale doshi velez is excited about methods to turn data into actionable knowledge. Towards using batch reinforcement learning to identify treatment options in healthcare. An emerging paradigm for modeling families of related tasks, year. Finale doshi velez and been kim to appear springer series on challenges in machine learning.
Finale doshi velez, an assistant professor of computer science at harvard, whose core research includes machine. Apr 02, 2017 how humans will lose control of artificial intelligence. Finale doshivelez interpretability in machine learning. Paulson school of engineering and applied sciences seas, and na li, assistant professor of electrical engineering and applied mathematics at seas, have. The program presents the experience of practitioners working with applications in the. Citeseerx nonparametric bayesian policy priors for. May 23, 2017 with a growing interest in interpretability, there is an increasing need to characterize what exactly we mean by it and how to sensibly compare the interpretability of different approaches. Finale doshi velez assistant professor of computer science harvard john a. Professor of the practice of health and human rights, harvard t. All eventbrite tickets will be fully reimbursed through the original payment method beginning next week. The infinite partially observable markov decision process. Finale doshi velez unice yoo, cpb, determining metal selectivity imparted by variations in the nramp metalbinding site rachelle gaudet gerard porter, undeclared, mechanistic study of catalytic ch bond amination theodore betley yemile bazaldua flores, heb, the effects of the r4 regulatory enhancer of the gdf5 gene on the development of. Towards using batch reinforcement learning to identify.
From software engineering to formal methods and tools, and back. Make machine learning interpretability more rigorous data. Just about everyone wants to be happy, but we all know that being happy is easier said than done. Biostatisticsbiomedical informatics big data b3d seminar speaker. I joke that we call it nice people in machine learning, says finale doshivelez, a harvard computer science professor and wiml executive board member. In other cases, the content of the topic may be of intrinsic scientific interest. Artificial intelligence is more powerful than ever. Ml software in health care, and proposes a lifecylebased framework given the iterative nature of aiml algorithms, including prespecified safe harbor provisions for change and postapproval audits of accuracy. Her research in machine learning, computational statistics, and data science develops methods for turning data into actionable knowledge in domains such as healthcare. Outside of research, i also have industry experience with software engineering and data science.
Bayesian nonparametric methods for reinforcement learning in partially observable domains. Finale doshivelez is an assistant professor of computer science at harvard. After 60 days you will be unable to print or save until you purchase finale and receive a serial number and access to the full collection of garritan sounds. The goals for the model vary depending on the application. Humanintheloop interpretability prior 2018 representation balancing mdps for offpolicy policy evaluation 2018 robust and efficient transfer learning with hidden parameter markov decision processes 2017 mind the gap. Paulson school of engineering and applied sciences seas, was awarded the everett mendelsohn excellence in mentoring award by the graduate student council each year at the everett mendelsohn excellence in mentoring awards, the graduate student council recognizes the faculty members who. Lifecycle regulation of artificial intelligence and machine. Doshi velez, finale and kortz, mason and budish, ryan and bavitz, christopher and gershman, samuel j. Kodi archive and support file community software vintage software apk msdos cdrom software cdrom software library console living room software sites tucows software library shareware cdroms software capsules compilation cdrom images zx spectrum doom level cd. Towards using batch reinforcement learning to identify treatment. Finale doshivelez is an assistant professor in computer science at the harvard paulson school of engineering and applied sciences. Topic modeling is a powerful tool for uncovering latent structure in many domains, including medicine, finance, and vision. Isaac lage andrew ross samuel j gershman been kim finale doshivelez. As machine learning systems become ubiquitous, there has been a surge of interest in interpretable machine learning.
Doshi velez will discuss innovations in her research group to apply these paradigms to real problems in healthcare. Semantic scholar profile for finale doshi velez, with 203 highly influential citations and 98 scientific research papers. Latent feature models are a powerful tool for modeling data with globallyshared features. Finale doshivelez is excited about methods to turn data into actionable knowledge. Finale doshi velez been kim as machine learning systems become ubiquitous, there has been a surge of interest in interpretable machine learning. Big data in the assessment of pediatric medication safety. Sign up for your own profile on github, the best place to host code, manage projects, and build software alongside 40 million developers. While youre focused on this, a bunch of small things go unnoticed, says dr. This cited by count includes citations to the following articles in scholar.
211 1161 345 1041 1335 1169 465 59 20 1377 24 585 1069 779 209 17 529 857 911 348 740 552 617 540 828 850 1162 655 1198