Artificial Intelligence for Social Impact: Learning and Planning in the Data-to-Deployment Pipeline
With the maturing of artificial intelligence (AI) and multiagent systems research, we have a tremendous opportunity to direct these advances toward addressing complex societal problems. In pursuit of this goal of AI for social impact, we as AI researchers must go beyond improvements in computational methodology; it is important to step out in the field to demonstrate social impact. To this end, we focus on the problems of public safety and security, wildlife conservation, and public health in low-resource communities, and present research advances in multiagent systems to address one key cross-cutting challenge: how to effectively deploy our limited intervention resources in these problem domains. We present case studies from our deployments around the world as well as lessons learned that we hope are of use to researchers who are interested in AI for social impact. In pushing this research agenda, we believe AI can indeed play an important role in fighting social injustice and improving society.
Contributors
Andrew Perrault
Harvard University
Fei Fang
Carnegie Mellon University
Arunesh Sinha
Singapore Management University
Milind Tambe
Harvard University
Evolution of a Robust AI System: A Case Study of AAAI’s AI-Alert
Since mid-2018, we have used a suite of artificial intelligence (AI) technologies to automatically generate the Association for the Advancement of Artificial Intelligence’s AI-Alert, a weekly email sent to all Association for the Advancement of Artificial Intelligence members and thousands of other subscribers. This alert contains ten news stories from around the web that focus on some aspect of AI, such as new AI inventions, AI’s use in various industries, and AI’s impacts in our daily lives. This alert was curated by-hand for a decade before we developed AI technology for automation, which we call “NewsFinder.” Recently, we redesigned this automation and ran a six-month experiment on user engagement to ensure the new approach was successful. This article documents our design considerations and requirements, our implementation (which involves web crawling, document classification, and a genetic algorithm for story selection), and our reflections after a year and a half since deploying this technology.
Contributors
Joshua Eckroth
i2k Connect
From ‘F’ to ‘A’ on the N.Y. Regents Science Exams: An Overview of the Aristo Project
AI has achieved remarkable mastery over games such as Chess, Go, and Poker, and even Jeopardy!, but the rich variety of standardized exams has remained a landmark challenge. Even as recently as 2016, the best AI system could achieve merely 59.3 percent on an 8th grade science exam. This article reports success on the Grade 8 New York Regents Science Exam, where for the first time a system scores more than 90 percent on the exam’s nondiagram, multiple choice (NDMC) questions. In addition, our Aristo system, building upon the success of recent language models, exceeded 83 percent on the corresponding Grade 12 Science Exam NDMC questions. The results, on unseen test questions, are robust across different test years and different variations of this kind of test. They demonstrate that modern natural language processing methods can result in mastery on this task. While not a full solution to general question-answering (the questions are limited to 8th grade multiple-choice science) it represents a significant milestone for the field.
Contributors
Peter Clark
Allen Institute for AI
Oren Etzioni
Allen Institute for AI
Tushar Khot
Allen Institute for AI
Daniel Khashabi
Allen Institute for AI
Bhavana Dalvi Mishra
Allen Institute for AI
Kyle Richardson
Allen Institute for AI
Ashish Sabharwal
Allen Institute for AI
Carissa Schoenick
Allen Institute for AI
Carissa Schoenick
Allen Institute for AI
Oyvind Tafjord
Allen Institute for AI
Niket Tandon
Allen Institute for AI
Sumithra Bhakthavatsalam
Allen Institute for AI
Dirk Groeneveld
Allen Institute for AI
Michal Guerquin
Allen Institute for AI
Michael Schmitz
Allen Institute for AI
Conversational Agents for Complex Collaborative Tasks
Dialogue is a very active area of research currently, both in developing new computational techniques for robust dialogue systems and in the active fielding of commercial conversational assistants such as Siri and Alexa. This paper argues that, while current techniques can be used to design effective dialogue-based systems for very simple tasks, they are unlikely to generalize to conversational interfaces that enhance human ability to solve complex tasks by interacting with AI reasoning and modeling systems. We explore some of the challenges of tackling such complex tasks and describe a dialogue model designed to meet these challenges. We illustrate our approach with examples of several implemented systems that use this framework.
Contributors
James Allen
Institute for Human and Machine Cognition, University of Rochester
Lucian Galescu
Institute for Human and Machine Cognition
Choh Man Teng
Institute for Human and Machine Cognition
Ian Perera
Institute for Human and Machine Cognition
Escaping the McNamara Fallacy: Towards more Impactful Recommender Systems Research
Recommender systems are among today’s most successful application areas of artificial intelligence. However, in the recommender systems research community, we have fallen prey to a McNamara fallacy to a worrying extent: In the majority of our research efforts, we rely almost exclusively on computational measures such as prediction accuracy, which are easier to make than applying other evaluation methods. However, it remains unclear whether small improvements in terms of such computational measures matter greatly and whether they lead us to better systems in practice. A paradigm shift in terms of our research culture and goals is therefore needed. We can no longer focus exclusively on abstract computational measures but must direct our attention to research questions that are more relevant and have more impact in the real world. In this work, we review the various ways of how recommender systems may create value; how they, positively or negatively, impact consumers, businesses, and the society; and how we can measure the resulting effects. Through our analyses, we identify a number of research gaps and propose ways of broadening and improving our methodology in a way that leads us to more impactful research in our field.
Contributors
Dietmar Jannach
University of Klagenfurt
Christine Bauer
Johannes Kepler University Linz
What Happens When AI Invents: Is the Invention Patentable?
In a decision published on April 27, 2020, the United States Patent and Trademark Office determined that only a human can be considered an inventor. So, who owns the patent when artificial intelligence makes the invention? What are the practical and logistical complications inherent in artificial intelligence-created inventions? With technology advancing faster than the legal system, what can we expect moving forward?
Contributors
Stan Gibson
Jessica Newman
The Association for the Advancement of Artificial Intelligence 2020 Workshop Program
The Association for the Advancement of Artificial Intelligence 2020 Workshop Program included twenty-three workshops covering a wide range of topics in artificial intelligence. This report contains the required reports, which were submitted by most, but not all, of the workshop chairs.
Contributors
Grace Bang
S&P Global
Guy Barash
Western Digital
Ryan Bea
University of Southampton
Jacques Cali
Blue Prism Group PLC
Mauricio Castillo-Effen
Lockheed Martin Advanced Technology Laboratories
Xin Cynthia Chen
University of Hong Kong
Niyati Chhaya
Adobe Research
Rachel Cummings
Georgia Institute of Technology
Rohan Dhoopar
Google Research
Sebastijan Dumanci
KU Leuven
Huáscar Espinoza
Commissariat à l´Énergie Atomique
Eitan Farchi
IBM Research
Ferdinando Fioretto
Syracuse University
Raquel Fuentetaja
Universidad Carlos III de Madrid
Christopher Geib
Smart Information Flow Technologies LLC
Odd Erik Gundersen
Norwegian University of Science and Technology
José Hernández-Orallo
Universitat Politècnica de València
Xiaowei Huang
University of Liverpool
Kokil Jaidka
National University of Singapore
Sarah Keren
Harvard University
Seokhwan Kim
Amazon Alexa AI
Michel Galley
Microsoft Research AI
Xiaomo Liu
S&P Global
Tyler Lu
Google Research
Zhiqiang Ma
S&P Global
Richard Mallah
Future of Life Institute
John McDermid
University of York
Martin Michalowski
University of Minnesota School of Nursing
Reuth Mirsky
University of Texas at Austin
Seán Ó hÉigeartaigh
Cambridge University
Deepak Ramachandran
Google Research
Javier Segovia-Aguas
Institut de Robòtica i Informàtica Industrial
Onn Shehory
Bar Ilan University
Arash Shaban-Nejad
University of Tennessee Health Science Center
Vered Shwartz
University of Washington
Siddharth Srivastava
Arizona State University
Kartik Talamadupula
IBM Research
Jian Tang
DiDi AI Labs
Pascal Van Hentenryck
Georgia Institute of Technology
Dell Zhang
Birkbeck, University of London
Jian Zhang
Microsoft Azure
What Happens When AI Invents: Is the Invention Patentable?
In a decision published on April 27, 2020, the United States Patent and Trademark Office determined that only a human can be considered an inventor. So, who owns the patent when artificial intelligence makes the invention? What are the practical and logistical complications inherent in artificial intelligence-created inventions? With technology advancing faster than the legal system, what can we expect moving forward?
Contributors
Stan Gibson
Jessica Newman
Reports of the Workshops Held at the 2020 International Association for the Advancement of Artificial Intelligence Conference on Web and Social Media
The workshop program of the Association for the Advancement of Artificial Intelligence’s Fourteenth International Conference on Web and Social Media was held June 8 to 20, 2020. The conference venue, which had originally been Atlanta, Georgia, USA, had to be revamped into a virtual edition, due to the 2019 Corona Virus Disease pandemic. There were five full-day workshops in the program. They included Emoji Understanding and Applications in Social Media (W1), Social Sensing: Narrative Analysis on Social Media (W2), News and Public Opinion (W3), Cyber Social Threats (W4), and Mediate: Social and News Media Misinformation (W5). This report contains summaries of all five workshops.
Contributors
Tarek Abdelzaher
University of Illinois at Urbana-Champaign
Jisun An
Hamad Bin Khalifa University
Marya Bazzi
Alan Turing Institute
Jeremy Blackburn
Binghamton University
Ugur Kursuncu
University of South Carolina
Yelena Mejova
Institute for Scientific Interchange (ISI) Foundation
Jérémie Rappaz
École Polytechnique Fédérale de Lausanne
Horacio Saggion
Universitat Pompeu Fabra
Panayiotis Smeros
École Polytechnique Fédérale de Lausanne (EPFL)
Sanjaya Wijeratne
oller Technologies, Inc.
Ning Yu
Leidos
Remembering Nils Nilsson
Former President of the Association for the Advancement of Artificial Intelligence and Stanford University professor Nils Nilsson died on Thursday, April 23, 2020 after a long illness. In this tribute, his former student Karen Myers provides personal reflections on Nilsson and the lasting impact he had on her life, as well as Nilsson’s contributions to the field of artificial intelligence.
Contributors
Karen Myers
SRI International