Annotated Table of Contents V B I P

AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014 AI Matters

Annotated Table of Contents Welcome to AI Matters ! Kiri Wagstaff, Editor Full article: http://doi.acm.org/10.1145/2639475.2639476 A welcome from the Editor of AI Matters and an en- couragement to submit for the next issue.

Artificial Intelligence: No Longer Just for O You and Me Yolanda Gil, SIGAI Chair Full article: http://doi.acm.org/10.1145/2639475.2639477 The Chair of SIGAI waxes enthusiastic about the cur- Drop-in Challenge games at RoboCup. rent state of and future prospects for AI developments V Peter Stone, Patrick MacAlpine, Katie Genter, and innovations. She also reports on high school and Sam Barrett. Full image and details: student projects featured at the 2014 Intel Science http://doi.acm.org/10.1145/2639475.2655756 and Engineering Fair. Using Agent-Based Modeling and Cul- Announcing the SIGAI Career Network I N and Conference tural Algorithms to Predict the Location Sanmay Das, Susan L. Epstein, and Yolanda of Submerged Ancient Occupational Gil Sites Full article: http://doi.acm.org/10.1145/2639475.2649581 Robert G. Reynolds, Areej Salaymeh, John O'Shea, and Ashley Lemke SIGAI has created a career networking website and Full article: http://doi.acm.org/10.1145/2639475.2639479 annual conference for the benefit of early career scientists. Benefits include mentoring, networking, and A collaboration between archaeologists and artificial job connections. intelligence experts has discovered ancient hunting sites submerged in over 120 feet of water in Lake Future Progress in Artificial Intelligence: Huron. This is the oldest known hunting ground in the world. P A Poll Among Experts Vincent C. Müller and Nick Bostrom A New Approach for Disruption Man- Full article: http://doi.acm.org/10.1145/2639475.2639478 B agement in Airline Operations Control When will intelligent systems surpass human intelli- Antonio J. M. Castro, Ana Paula Rocha, and gence? This study surveyed experts and found that Eugénio Oliveira they predict that this time, sometimes referred to as the singularity, will occur before 2080. The study also Full article: http://doi.acm.org/10.1145/2639475.2639480 found that nearly one third of experts surveyed have This new book describes the application of a multi- strong concerns about the negative impact on human- agent approach to address challenges in airline op- ity. erations. It provides rapid responses to disruptive events so as to minimize the impacts on the crew and passengers.

ISSN 2372-3483 1 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014

The NY AI Summit: A Meeting of AI Dis- E  cipline Leaders   Organized by IJCAI and AAAI   Francesca Rossi (IJCAI President) and Manu-  ela Veloso (AAAI President) Full article: http://doi.acm.org/10.1145/2639475.2639481

AAAI and IJCAI co-organized a meeting to discuss   the future of AI, including conference coordination,  how AI sub-disciplines relate, and societal impact. This report features highlights of the event and describes next steps to better coordinate sub-disciplines and create an open information structure to dissemi- nate and coordinate community-wide information. Information network for the 2011 Fukushima Submit your Ph.D. brieﬁng here! D V earthquake. Jure Leskovec and Manuel See the AI Matters website for more info. Gomez Rodriguez. Full image and details: http://doi.acm.org/10.1145/2639475.2655757 Upcoming Conferences Registration discount for SIGAI members. AI Matters Editorial Board WI-IAT ‘14: Web Intelligence and Intelligent Kiri Wagstaff, Editor-in-Chief, JPL/Caltech Agent Technology. Warsaw, Poland, Aug. 11- Sanmay Das, Washington Univ. of Saint Louis 14, 2014. Alexei Efros, Univ. of CA Berkeley ASE ‘14: ACM/IEEE International Conf. on Susan L. Epstein, The City Univ. of NY Automated Software Engineering. Vasteras, Yolanda Gil, ISI/Univ. of Southern California Sweden, Sept. 15-19, 2014. Doug Lange, U.S. Navy RecSys ‘14: ACM Conf. on Recommender Sys- Xiaojin (Jerry) Zhu, Univ. of WI Madison tems. Foster City, CA. Oct. 6-10, 2014. AAAI Doctoral Consortium ’15. Austin, TX. Submissions Jan. 25-25, 2015. (Submission: Sept. 22, 2014) B Book Announcement HRI ‘15: ACM/IEEE International Conf. on Human-Robot Interaction. Portland, OR. Mar. D Ph.D. Dissertation Brieﬁng 2-5, 2015. Event Report (Submission: Oct. 3, 2014) E IUI ‘15: International Conf. on Intelligent User I AI Impact Interfaces. Atlanta, GA. Mar. 29 - Apr. 1, 2015. (Submission: Oct. 17, 2014) N AI News O Opinion Links SIGAI website: http://sigai.acm.org/ P Paper Précis Newsletter: http://sigai.acm.org/aimatters/ Video or Image Twitter: http://twitter.com/acm_sigai/ V Details at http://sigai.acm.org/aimatters Edition DOI: 10.1145/2639475

2 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014

Welcome to AI Matters Kiri Wagstaff, Editor (Jet Propulsion Laboratory, California Institute of Technology; [email protected]) DOI: 10.1145/2639475.263947

Welcome to AI Matters, the new quarterly news- ing a recently published poll of what AI experts letter for SIGAI, the ACM Special Interest Group think about the evolution of AI, how AI methods on Artificial Intelligence. This newsletter features help underwater archaeology, AI methods for air- ideas and announcements of interest to the AI line operations, a report on the NY AI Summit, community. These include: and an announcement about the newly created Book Announcement: Description of a newly SIGAI Career Network and conference. published book and its major contributions. Dissertation briefings: Extended abstracts We encourage you to submit your own material from new Ph.D.s. for future issues. You can learn more about Event reports: Technical conference or work- submissions at the AI Matters website, where shop reports, policy forums, or community you can also download submission templates: events on topics of general interest to an AI http://sigai.acm.org/aimatters/ Authors retain audience. copyright to their contributions, which are pub- AI Impact: Description of an AI system or lished by the ACM Digital Library. Submissions method that has had a tangible impact on the are reviewed by the AI Matters Editorial Board. world outside of the AI research community. AI News: Innovations, open source AI soft- We hope you enjoy this newsletter and find that ware, course materials, challenges and competi- it points you in new directions or encourages tions, and other news of broad interest to AI re- new ideas and innovation. searchers and practitioners. Opinion: Discussion of thought-provoking issues and responses to previous items. Kiri Wagstaff is the Editor Paper Précis: Short summary of the major of AI Matters. She is also contributions of a recently published AI paper, a senior researcher in written for the general AI audience. machine learning and Tutorial: Short introduction or explanation of data analysis at the Jet an AI concept or technique. Propulsion Laboratory in Videos and Images: Audio-visual materials Pasadena, CA. She with content of general interest to an AI audi- serves as a tactical plan- ence. ner for the Mars Explora- tion Rover Opportunity In this debut issue, we begin with an enthusiastic and continually brainstorms ways to make the discussion by the Chair of SIGAI of the broad rover more autonomous. relevance of AI. We also include pieces discuss-

3 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014

O Artiﬁcial Intelligence: No Longer Just for You and Me Yolanda Gil, SIGAI Chair (Information Sciences Institute and Department of Computer Science, University of Southern California; [email protected]) DOI: 10.1145/2639475.2639477

As Chair of SIGAI, I wanted to share the excitement that I see emerging in our ﬁeld for this ﬁrst issue of AI Matters.

First, AI is having an impact in the world and can no longer be considered an exotic boutique research area. A wide range of AI technologies are permeating industry, science, entertainment, and our everyday lives. From the Siri speech-based phone assistant, to IBM’s Watson learning from text to become a Jeopardy game winner, to self- driving cars, AI is becoming directly present in people’s lives. People have come to appreciate the potential of intelligent machines in many areas of societal relevance. The rising challenges Figure 1. A rising tide of students interested in AI: of big data and data science cannot be met with- More than two thirds of the computer science posters out AI playing a major role not only in mining but presented at the 2014 ISEF were on AI topics such also in understanding, summarizing, and model- as machine learning, robotics, and image processing. ing data. The Google Knowledge Graph has made knowledge bases familiar to everyone, and of the student posters focused on biomedical the Wikidata project at the Wikimedia Foundation applications of AI. In addition to those CS post- has tens of thousands of contributors building a ers, we found thirty or so more from other areas semantic network version of Wikipedia that had of engineering and science that were relevant to accumulated 30M statements after just one year. AI. That signified around one hundred AI posters The Web is becoming increasingly structured of excellent quality that made judging really chal- with hundreds of knowledge bases and ontolo- lenging. gies that are beginning to change how we access and interpret information. This is a truly ex- Our top award went to an agent-based simula- citing time for our field. tion for understanding the spread of disease. Our second award went to a computer vision al- Another major reason for great excitement is the gorithm for grading the stage of prostate cancer. enthusiasm for AI that is palpable in new genera- Our third award recipient, who ended up taking tions. I will recount here my recent experience also the top award at the fair, used machine as a judge for high school student AI projects, learning to analyze how gene mutations affect already selected among the best in the world. the properties of proteins. Many of these stu- This was at the annual international Intel Science dents had formulated and carried out their pro- and Engineering Fair (ISEF) (which used to be jects independently, just researching about AI on the Westinghouse Science Fair). I was extremely the Web. Their excitement was very palpable. impressed with the large amount of students in- One student told me his hobby was to read AI terested in AI, the quality of the projects, and the papers from the sixties. Another student in the excitement of the students about our field. Of biomedical engineering area overheard me say the hundred or so CS posters, two-thirds were that I was there to judge AI projects and ap- on AI. The most popular topics were machine proached me to tell me he had enjoyed a lot the learning, robotics, and image processing. Many AI work that he had done in his project and

4 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014 wanted to learn more about how to get involved. standing of the brain, and form new areas of sci- Students from countries like Nigeria, Georgia, ence such as social computation and the Se- Peru, Oman, and many others represented the mantic Web. talent of this new generation. The future of our field is in great hands. SIGAI is committed to helping our community grow. Its membership is diverse and includes Finally, an exciting recent development is the not only researchers and students but also in- announcement of the XPRIZE for Artificial Intelli- dustry and government practitioners. SIGAI has gence jointly with TED. The challenge is to put embarked on new activities that are geared to an AI system on the TED stage to give a talk that grow and strengthen our field. SIGAI officers will get a standing ovation. Addressing this chal- work with ACM’s committees and initiatives that lenge would require fundamental advances in are reaching out to new constituencies like CS many areas of AI research. But that is not a new teachers, garage tinkerers, policy makers, and thing, for example we have had the Turing test the international community. Please contact any as a standing challenge for decades and many SIGAI officer if you are interested in being part of other challenges with awards. What is notable any of our community building efforts. about the A.I. XPRIZE is the crowdsourcing of the rules that will test how the AI system demon- strates intelligence. There is some chance that, Yolanda Gil was re-elected as has happened with other similar challenges, Chair of ACM SIGAI in some students or perhaps garage tinkerers will 2013. She is Director of pull together a competitive entry, even a winning Knowledge Technologies one. at the Information Sci- ences Institute and Re- The future of our field is bright. The trends above search Professor of Com- suggest that we need to broaden our activities puter Science at the Uni- and reach practitioners, adopters, and students versity of Southern Califor- beyond the arena of academic research. We nia. She is a Fellow of need to get the public interested when there are AAAI. Her research inter- major breakthroughs in our field. Astronomers, ests include intelligent user interfaces, biologists, and physicists do it – why shouldn’t knowledge-rich problem solving, semantic work- we? Our quest is important and we must get flows, AI-mediated scientific collaboration, others excited, as we bring to the world smart provenance, and semantic web. machines like no others, improve our under-

5 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014

N Announcing the SIGAI Early Career Researchers Network and Conference Sanmay Das (Washington University in St. Louis; [email protected]) Susan L. Epstein (Hunter College and The Graduate Center of the City University of New York; [email protected]) Yolanda Gil (Information Sciences Institute and Department of Computer Science, University of Southern California; [email protected]) DOI: 10.1145/2639475.2649581

Any research field is as healthy as the new talent changes through the SIGAI Career Network that it is able to attract, and AI is no exception. website. For this reason, AI conferences hold mentoring events for doctoral students and researchers in SIGAI CNC will feature presentations from stu- the early stages of their careers to support their dents who have recently completed (or nearly advancement and connections to other re- completed) their dissertations. Applicants will be searchers in the field. SIGAI holds one such Ph.D. students who are about to defend and cur- event annually at the AAAI conference: the AAAI/ rent postdocs. To apply, a researcher will submit SIGAI Doctoral Consortium. But we think that a CV, a research statement, and letters of rec- much more can be done, as these events are ommendation. Based only on research quality, held once a year and do not necessarily cover all several applicants will be selected (by an inde- the topics that young researchers would want to. pendent panel or program committee) and invited to give an oral presentation (20-25 minute) To support these goals, SIGAI is planning to and/or a poster presentation. Each presentation launch a Career Network website and an associ- will be a broad summary of their thesis or post- ated annual conference. Our goal is to create a graduate research, rather than a single paper. network for early-career scientists, one that will support them as they transition from Ph.D. / SIGAI will contribute significant travel funding for postdoctoral research to independent research in many of the selected students. Registration at academia, industry, or government. The SIGAI CNC will be open to all SIGAI members, with a Career Network Conference (SIGAI CNC) will be token fee for any graduate student attendees. an official ACM conference that showcases the The event’s format will be designed with each work of early career researchers to their potential year’s event chairs. Accepted submissions will mentors and employers. This showcase will be be published in the ACM Digital Library and dis- a significant extension beyond what currently seminated through the Career Network website. occurs at AI conferences. In 2015, we plan to hold CNC in Austin, Texas, collocated with AAAI. SIGAI CNC will also include networking opportu- In parallel with the conference, the Career Net- nities in the form of interactive poster sessions, work website will provide a virtual community for professional booths, mentoring events, and a job AI researchers in the early stages of their ca- fair. One of the main goals is to allow young re- reers. searchers to network with researchers outside of academia. The experience of most Ph.D. stu- SIGAI CNC dents and postdocs is limited to the academic SIGAI will hold an annual conference, SIGAI world. SIGAI believes that the opportunity to CNC, to showcase high-quality research from meet and interact with the research community graduating Ph.D.s and postdocs. CNC will also in industry and government could broaden early- include a wide range of opportunities for career career scientists’ horizons, and prepare them for development and mentoring. CNC will be a face- future careers outside of academia. to-face event complemented by on-line ex-

6 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014

The Career Network Website process, employers have first-round interviews To facilitate the creation of a virtual community with candidates at an annual meeting or conven- for early-career AI researchers (those who have tion in the fall or winter. Moreover, these inter- completed their Ph.D.s within the last six years, views cost little, because both employers and job or graduate students in the final year of their seekers already attend the annual meeting; the Ph.D. program), SIGAI will launch the SIGAI Ca- main issues are time and scheduling. First-round reer Network website in the fall of 2014. The interviews serve both employers and job seekers website will be run by early-career researchers well. Employers can briefly screen candidates under oversight from SIGAI. without an on-campus or on-site visit, while job seekers can establish contact with employers The SIGAI Career Network website will not only and test their potential fit with them before more connect early-career AI researchers, but also substantial on-site interviews. This gives job provide a matching service between potential seekers an early idea about work possibilities employers and recent Ph.D. graduates. Recent and a better overall perspective on their job Ph.D. graduates and other early-career research. Overall, there are fewer failed searches searchers, as well as potential employers, can and better matches. For more on this issue, see register to make use of the website. Information this blog post by Lance Fortnow: on potential employers would be publicly avail- http://blog.computationalcomplexity.org/2007/02/ able (simply, University X Dept. Y, or Company Z organizing-academic-job-market.html . seeks to hire in AI). Potential employees either make their profiles public or restrict them only to While we envision SIGAI CNC as an exciting op- potential employers they select. The latter would portunity to gather the best young researchers in support personal privacy, for example, for some- AI in a forum where the entire community can one seeking a new job. learn about their research, it also presents opportunities to connect job seekers with potential SIGAI CNC and the Career Network website will employers. The conference will be well timed (in complement each other to provide a community January) for both job seekers and employers. for support, information sharing, and networking SIGAI CNC will provide an important service to among early-career AI researchers. our community.

On the “Job Market” Aspects of the Career SIGAI and AAAI Collocation Network and CNC AAAI and SIGAI already cooperate with the Many computer scientists are frustrated by how AAAI/SIGAI Doctoral Consortium (DC). SIGAI disorganized our job market is in comparison CNC and the DC will be complementary events: those of other disciplines. In particular, there is DC will focus on students at early stages of their limited information on the range and nature of PhD and at institutions without many faculty in the many non-academic jobs available to gradu- AI, while CNC will focus on soon-to-graduate ating AI Ph.Ds. These jobs exist in government PhD students and post-doctoral researchers. labs, at research organizations that do govern- SIGAI CNC will be held immediately before the ment contract work, and at smaller industry- main AAAI conference, in parallel with the work- research labs and startups. There are also some shops and the DC. little-known teaching opportunities in predomi- nantly undergraduate institutions and smaller Summary colleges. SIGAI’s planned activities for early career AI researchers and AAAI’s move to a winter confer- Most academic disciplines pursue a more coor- ence schedule have presented a rare opportunity dinated approach to hiring, even when significant for AI and for our organizations: the collocation of options are available outside academia (in, for SIGAI CNC with the annual AAAI meeting. This example, economics and finance). In the typical will benefit not only young researchers, who will

7 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014 showcase their work and get career advice, but and roboticists to identify important principles also potential employers, given the event’s tim- about knowledge and learning, and to help coming. SIGAI CNC will become a destination for AI puters exploit them. Her current research inter- scientists to discuss the best new research and ests include plausible recommendations, human- meet the people who make it possible. multi-robot teams for search and rescue, protein- protein interaction networks, and parallel search For the most up-to-date information on the SIGAI for solutions to constraint satisfaction problems. Career Network, see: She is Professor of Computer Science at Hunter http://sigai.acm.org/cnc/ College and The Graduate Center of The City University of New York. Sanmay Das’ research interests are in multi-agent Yolanda Gil was re-elected systems, machine learn- Chair of ACM SIGAI in ing, and computational so- 2013. She is Director of cial science. He is the Knowledge Technologies at vice-chair of SIGAI. the Information Sciences Institute and Research Pro- fessor of Computer Sci- ence at the University of Southern California. She is a Fellow of AAAI. Her re- Susan L. Epstein develops search interests include knowledge representations intelligent user interfaces, knowledge-rich prob- and machine learning al- lem solving, semantic workﬂows, AI-mediated gorithms to support pro- scientiﬁc collaboration, provenance, and seman- grams that learn to be ex- tic web. perts. An interdisciplinary scholar, she has worked with and published for mathematicians, psy- chologists, geographers, linguists, microbiologists,

8 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014

P Future Progress in Artiﬁcial Intelligence: A Poll Among Experts Vincent C. Müller (Future of Humanity Institute, University of Oxford & Anatolia College/ACT; [email protected]) Nick Bostrom (Future of Humanity Institute, Oxford University; [email protected]) DOI: 10.1145/2639475.2639478

This is an abbreviated version of: Müller, Vincent formance of humans in virtually all domains of C. and Bostrom, Nick (forthcoming 2014), ‘Fu- interest.” (Bostrom, 2014 ch. 2). ture progress in artificial intelligence: A poll among experts’, in Vincent C. Müller (ed.), Fun- For the questionnaire we settled for a definition damental Issues of Artificial Intelligence (Syn- that a) is based on behavioral ability, b) avoids these Library; Berlin: Springer). A pre-print of the the notion of a general ‘human–level’ and c) uses full paper is available on a newly coined term. We put this definition in the http://www.sophia.de/publications.htm. Please preamble of the questionnaire: “Define a ‘high– cite the full version. level machine intelligence’ (HLMI) as one that can carry out most human professions at least Abstract: In some quarters, there is intense as well as a typical human.” concern about high–level machine intelligence and superintelligent AI coming up in a few dec- 2. Questionnaire ades, bringing with it significant risks for human- The questionnaire was carried out online by invi- ity; in other quarters, these issues are ignored or tation to particular individuals from four different considered science fiction. We wanted to clarify groups. The groups we asked were: what the distribution of opinions actually is, what • PT–AI: Participants of the conference on “Phi- probability the best experts currently assign to losophy and Theory of AI”, Thessaloniki Octo- high–level machine intelligence coming up within ber 2011, organized by one of us (see Müller, a particular time–frame, which risks they see 2012, 2013). Response rate 49%, 43 out of 88. with that development and how fast they see • AGI: Participants of the conferences of “Artifi- these developing. We thus designed a brief cial General Intelligence” (AGI 12) and “Im- questionnaire and distributed it to four groups of pacts and Risks of Artificial General Intelli- experts. Overall, the results show an agreement gence” (AGI Impacts 2012), both Oxford De- among experts that AI systems will probably cember 2012, organized by both of us (see reach overall human ability around 2040-2050 Müller, 2014). Response rate 65%, 72 out of and move on to superintelligence in less than 30 111. years thereafter. The experts say the probability • EETN: Members of the Greek Association for is about one in three that this development turns Artificial Intelligence (EETN). Response rate out to be ‘bad’ or ‘extremely bad’ for humanity. 10%, 26 out of 250 (asked via e-mail list). • TOP100: The 100 ‘Top authors in artificial intel- 1. Problem ligence’ by ‘citation’ in ‘all years’ according to The idea of the generally intelligent agent con- Microsoft Academic Search in May 2013. Re- tinues to play an important unifying role for the sponse rate 29%, 29 out of 100. discipline(s) of artificial intelligence, it also leads Total response rate: 31%; 170 out of 549. We fairly naturally to the possibility of a super- also review prior work in (Michie, 1973, p. 511f), intelligence. If we humans could create artificial (Moor, 2006), (Baum, Goertzel, & Goertzel, general intelligent ability at a roughly human 2011): and (Sandberg & Bostrom, 2011). level, then this creation could, in turn, create yet higher intelligence, which could, in turn, create 3. Answers yet higher intelligence, and so on … “We can 1) “In your opinion, what are the research ap- tentatively define a superintelligence as any in- proaches that might contribute the most to the tellect that greatly exceeds the cognitive per- development of such HLMI?: …” There were

9 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014

no significant differences between groups Experts allocate a low probability for a fast take- here, except that ‘Whole brain emulation’ got off, but a significant probability for superintelli- 0% in TOP100, but 46% in AGI. gence within 30 years after HLMI. 2) “For the purposes of this question, assume that human scientific activity continues without 4) For the overall impact of superintelligence on major negative disruption. By what year would humanity, the assessment was: you see a (10%/50%/90%) probability for % PT-AI AGI EETN TOP ALL such HLMI to exist?” 100 Predicted years, sorted by HLMI probability: Extremely good 17 28 31 20 24 10% Median Mean St. Dev. On balance good 24 25 30 40 28 More or less PT-AI 2023 2043 81 neutral AGI 2022 2033 60 23 12 20 19 17 On balance bad EETN 2020 2033 29 17 12 13 13 13 Extremely bad TOP100 2024 2034 33 (existential ALL 2022 2036 59 catastrophe) 50% Median Mean St. Dev. 18 24 6 8 18 2048 2092 166 PT-AI We complement this paper with a small site on AGI 2040 2073 144 http://www.pt-ai.org/ai-polls/. On this site, we EETN 2050 2097 200 provide a) the raw data from our results, b) the TOP100 2050 2072 110 basic results of the questionnaire, c) the com- ments made, and d) the questionnaire in an on- ALL 2040 2081 153 line format where anyone can fill it in. 90% Median Mean St. Dev. PT-AI 2080 2247 515 References AGI 2065 2130 202 Baum, S. D., Goertzel, B., & Goertzel, T. G. EETN 2093 2292 675 (2011). How long until human-level AI? Results from an expert assessment. Technological TOP100 2070 2168 342 Forecasting & Social Change, 78(1), 185-195. ALL 2075 2183 396 Bostrom, N. (2014). Superintelligence: Paths, dangers, strategies. Oxford: Oxford University The median is 2050 or 2048 for three groups and Press. 2040 for AGI – a relatively small group that is Michie, D. (1973). Machines and the theory of defined by a belief in early HLMI. We would sug- intelligence. Nature, 241(23.02.1973), 507- gest that a fair representation of the result in 512. non–technical terms is: Experts expect that be- Moor, J. H. (2006). The Dartmouth College artifi- tween 2040 and 2050 high–level machine intelli- cial intelligence conference: The next fifty gence will be more likely than not. years. AI Magazine, 27(4), 87-91. Müller, V. C. (Ed.). (2012). Theory and philoso- 3) For the transition from HLMI to superintelli- phy of AI (Vol. 22/2 - Special volume): Sprin- gence, responses were: ger. Müller, V. C. (Ed.). (2013). Theory and philoso- Median Mean St. Dev. phy of AI (Vol. 5). Berlin: Springer. Within 2 years 10% 19% 24 Müller, V. C. (Ed.). (2014). Risks of artificial general intelligence (Vol. 26/3 - Special Volume): 75% 62% 35 Within 30 years Taylor & Francis.

10 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014

Sandberg, A., & Bostrom, N. (2011). Machine has published a number of articles and edited intelligence survey. FHI Technical Report, volumes on the philosophy of computing, the phi- 2011(1). Available from losophy of AI and cognitive science, the philoso- http://www.fhi.ox.ac.uk/research/publications/ phy of language, and related areas. He works at Anatolia College/ACT and at the University of Oxford. Vincent C. Müller's research focuses on the Nick Bostrom is a profes- nature and future of com- sor of the Philosophy & putational systems, par- Oxford Martin School, Di- ticularly on the prospects rector of the Future of of artiﬁcial intelligence. Humanity Institute, and He is the coordinator of Director of the Pro- the European Network for gramme on the Impacts of Cognitive Systems, Ro- Future Technology at the botics and Interaction University of Oxford. (2009-2014) with over 900 members (3.9 mil. €, www.eucognition.org). He has organized a number of prominent conferences in the ﬁeld. Müller

11 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014

I Using Agent-Based Modeling and Cultural Algorithms to Predict the Lo- cation of Submerged Ancient Occupational Sites Robert G. Reynolds (Wayne State University, University of Michigan Ann Arbor; [email protected]) Areej Salaymeh (Wayne State University) John O'Shea (University of Michigan Ann Arbor) Ashley Lemke (University of Michigan Ann Arbor) DOI: 10.1145/2639475.2639479

Some of the most pivotal questions in human history, such as the origins of early human cul- ture, the spread of hominids out of Africa, and the colonization of New World necessitate the investigation of archaeological sites that are now under water. These contexts have unique poten- tials for preserving ancient sites without distur- bance from later human occupation. The Alpena- Amberley Ridge (AAR) beneath modern Lake Huron in the North American Great Lakes offers unique evidence of prehistoric caribou hunters for a time period that is very poorly known on land.

An NSF funded research team headed by Ar- chaeologist John O’Shea from the University of Michigan, Guy Meadows an Engineer from the University of Michigan, and Robert Reynolds from Wayne State University have developed a novel approach to predicting the location of ancient hunting sites in over 120 feet of water un- derneath Lake Huron using techniques from Arti- ﬁcial Intelligence.

In addition to the archaeological investigations, intelligent systems was employed to better un- Figure 1. Predicted Annual Spring and Fall Migration derstand the movement of caribou and caribou routes using the intelligent agent model of caribou hunters on the AAR. Drawing on the environ- herd movement. mental reconstruction and a detailed map produced from side scan and multi-beam sonars, an likely to contain hunting structures using the intelligent agent based simulation of caribou herd caribou herd movement simulation data and eth- movement across the AAR was developed (Rey- nographic information (Reynolds, 1999). An im- nolds et al., 2013; Vitale et al., 2011). This simu- portant result of the simulation was the prediction lation provided a level of social intelligence to the that there should be distinctive routes for the individual animals as they iteratively transited autumn and spring migrations (Figure 1). The and learned the landscape over time. simulation also highlighted two critical choke points within the study area where all preferred A machine learning tool, Cultural Algorithms, migrations routes for both seasons converge. based upon models of Cultural Evolution gener- Drop 45 is located at one of these predicted ated “hot spots” representing areas that were choke points.

12 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014

The newly discovered Drop 45 Drive Lane is the Dr. Robert G. Reynolds is most complex hunting structure found to date, a Professor in the De- and it provides a compelling picture of later partment of Computer Paleoindian/Early Archaic caribou hunting in the Science in the College of Great Lakes region (O’Shea et al., 2014). The Engineering at Wayne site also provides important insight into the social State University and a and seasonal organization of these early Great Visiting Associate Re- Lakes inhabitants. When combined with envi- search Scientist in the ronmental and simulation studies, it is suggested Museum of Anthropologi- that distinctly different seasonal approaches cal Archaeology at the were used by early hunters on the AAR, with University of Michigan-Ann Arbor, Michigan. He autumn hunting being carried out by small received his Ph.D. degree in Computer Science, groups and spring hunts being conducted by specializing in Artificial Intelligence, in 1979 from larger collective groups. the University of Michigan, Ann Arbor. He is currently a professor of Computer Science and di- The Drop 45 Drive Lane and associated artifacts rector of the Artificial Intelligence Laboratory at are the oldest known evidence of ancient hunting Wayne State University. Dr. Reynolds produced structures in the world. As such, they provide an a framework, Cultural Algorithms, in which to ex- unprecedented insight into the social and sea- press and computationally test various theories sonal organization of early peoples in the Great of social evolution using multi-agent simulation Lakes Region as well Paleo-Indians in general. models. He has applied these techniques to In addition, the interdisciplinary research pro- problems concerning the origins of the state in gram provides a general model for the investiga- the Valley of Oaxaca, Mexico, the emergence of tion of submerged prehistoric landscapes using ancient urban centers, the origins agriculture, Artificial Intelligence techniques (O’Shea et al., and the disappearance of the Ancient Anasazi in 2014). Southwestern Colorado.

References John O'Shea is the Emer- Reynolds, R., Vitale, K., Che, X., O’Shea, J., son F. Greenman Colle- & Salaymeh, A. (2013). Using virtual worlds to giate Professor of Anthro- facilitate the exploration of ancient landscapes. pology at the University of International Journal of Swarm Intelligence Re- Michigan and Curator of search, 4(2), 49-83. Great Lakes Archaeology Vitale, K., Reynolds, R., O’Shea, J., & Meadows, at the Museum of Anthro- G. (2011). Exploring ancient landscapes under pology, University of Lake Huron using Cultural Algorithms. Procedia Michigan. He received his Computer Science, 6, 303-310. Ph.D. in Prehistoric Ar- Reynolds, R. G. (1999). in New Ideas in Optimi- chaeology at Cambridge zation, eds. Corne, D., Glover, F., & Dorigo, M. University in 1979. His (McGraw Hill Press), 367-378. work has focused on using archaeology to O’Shea, J., Lemke, A.K., Sonnenberg, E., Rey- document tribal societies before they are dis- nolds, R. G., & Abbot, B. (2014). A 9000-year- rupted by contact with modern states and em- old caribou hunting structure beneath Lake pires. His research areas include Native North Huron. Proceedings of the National Academy of America on the eve of European Contact and on Science, April 28, 2014. the development of Bronze Age societies in the Carpathian Basin and their larger connections into Europe and the Mediterranean. Professor O’Shea is also the director of the Museum’s Ar- chaeology Underwater program which considers

13 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014 both historic era shipwrecks and submerged prehistoric settlements in the Great Lakes. Areej Salaymeh is a Ph.D. candidate student Ashley Lemke is a doc- in computer science de- toral candidate in the department at Wayne State partment of Anthropology University. She is a mem- and Museum of Anthropo- ber of the Artificial Intelli- logical Archaeology at the gence Lab. Her research University of Michigan. interest is Artificial Intelli- Her primary research in- gence, Game program- terests include the an- ming, Artificial Intelligence thropology of hunting and in Game, Evolutionary archaeology of hunter- Algorithms, Cultural Algo- gatherers. These re- rithms, Multi-Objective Evolutionary Algorithms. search questions have led her to work in North America and Europe on both terrestrial and underwater archaeological projects from the Lower Paleolithic to 20th century shipwrecks.

14 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014

B A New Approach for Disruption Management in Airline Operations Control Antonio J. M. Castro (LIACC, University of Porto; [email protected]) Ana Paula Rocha (LIACC, FEUP, University of Porto; [email protected]) Eugénio Oliveira (LIACC, FEUP, University of Porto; [email protected]) DOI: 10.1145/2639475.2639480

As a sequel of a recent Ph.D. thesis at the Uni- optimization methods. However, nowadays, with versity of Porto and LIACC research lab, a new the increasingly complex and huge traffic at air- book appeared at Springer, Series: Studies in ports, the real challenge is how to react to unex- Computational Intelligence, Vol. 562, 2014, XII, pected events that may cause plan disruptions, 242 p., entitled A New Approach for Disruption leading to flight delays. Management in Airline Operations Control by An- tonio J. M. Castro, Ana Paula Rocha, and Eugé- Moreover these disruptive events usually affect nio Oliveira. at least three different dimensions of the situa- tion: the aircraft assigned to the flight, the crew The relevance of this book is mainly to show how assignment and, often forgotten, the passengers’ the multi-agent system paradigm can be used to journey and satisfaction. solve a very relevant real life real size problem. The book introduces a new concept for disrup- This book includes answers to this challenge and tion management in current Airline Operations proposes the use of the Multi-agent System Control Centers, taking into account their organi- paradigm to rapidly compose a multi-faceted so- zation, tools, problems, methods and costs. lution to the disruptive event taking into consid- Most of the research efforts dealing with airline eration possible preferences of those three key scheduling have been done through off-line plan aspects of the problem.

Figure 1. MASDIMA user interface displaying, on the left, problems encountered and proposed solutions, including costs and, on the right, ﬂights being monitored by the system. On top right we may access relevant information for each dimension (Aircraft, Crew, Pax) of selected ﬂights.

15 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014

Negotiation protocols taking place between Engineering, University of Porto. She partici- agents that are experts in solving the different pated in European as well as national funded problem dimensions (regarding aircrafts, crew projects concerning the use of intelligent agents and passengers), combination of different utility advanced features for applications. Her main functions and, not less important, the inclusion of current research topics of interest include Agent- the human in the automatic decision-making loop based frameworks for B2B, Multi-Agent Learn- make MASDIMA, the system described in this ing, Negotiation, Argumentation and Trust. She book, well suited for real-life plan-disruption was co-organizer of Artificial Intelligence and management applications. Multi Agent Systems related workshops. She is member of DAIAS (Distributed Artificial Intelli- Antonio Castro was born gence and Agent-based Simulation) group at LI- in 1965 in Porto, Portugal ACC (Laboratory of Artificial Intelligence and and studied at Porto Poly- Computer Science) since 1990. She is also technic Engineering Insti- member of APPIA (Portuguese Association for tute where he got his de- Artificial Intelligence). gree in Information Sys- tems Engineering in Eugénio Oliveira is full 1997. In 2007 he got his professor at the University Master Degree in Artificial of Porto, Faculty of Engi- Intelligence and Intelli- neering. He is director of gent Systems from the the Doctoral Program in Faculty of Engineering, Informatics Engineering University of Porto and the Ph.D. in Computer and Coordinator of the Engineering from the same institution in 2013. scientific research AI Lab Additionally, he has a postgraduate course in Air (LIACC). He got his Ph.D. Transport Operations from ISEC in 2008. Anto- in 1984 at the New Uni- nio works for TAP Portugal since 1990 and cur- versity of Lisbon in Logic rently he is a Board Advisor for IT/IS projects and P r o g r a m m i n g f o r responsible for projects related with airline op- Knowledge-based Sys- erations control. Antonio is also the CEO of tems. He was Guest Academic at IBM-IEC, La MASDIMA (http://www.masdima.com), a start-up Hulpe, in 84-85. He was, during several sab- company created together with two other col- batical terms, at the University of London leagues, from the research made during his (QMWC), University of Nice, University of Ph.D. where a Multi-Agent System for Disruption Utrecht, URJC (Madrid) and INPG (Grenoble). Management applied to Airline Operations Con- He was awarded with Gulbenkian Prize for Sci- trol was proposed, that includes the passenger ence and Technology. His main interests are on point of view in the Irregular Operations Man- Distributed Artificial Intelligence, Multi-agent Sys- agement Process (IROPS). tems, Trust And Reputation models and Text Mining. He supervised more than 20 Ph.D. stu- Ana Paula Rocha got her dents in these mentioned subjects. degree in Electrical and Computers Engineering at the University of Porto in 1990. She got her Ph.D. in Computer Engineering from the same institution in 2002. Currently, she is Auxiliary Professor at the Department of Informatics

16 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014

E The NY AI Summit: A Meeting of AI Discipline Leaders Francesca Rossi (President, IJCAI; University of Padova; [email protected]) Manuela Veloso (President, AAAI; Carnegie Mellon University; [email protected]) DOI: 10.1145/2639475.2639481

AI is becoming pervasive in our lives. Its impact associations of AI researchers, and the impact of on society is increasing every day. Its potential AI in the society at large. is enormous and there have recently been several outstanding achievements. The meeting was held in New York City on Feb- ruary 24-25, 2014, and saw the participation of However, AI as a science is very complex and about 50 of the most active researchers in AI, vast. Because of this, historically and inevitably coming from academia, companies, large re- it has fragmented into sub-disciplines, all with search centers, and AI-related organizations and fantastic contributions. It could however be the journals. The meeting was termed “The NY AI case now that the existence of a large number of Summit.” AI-related disciplines, conferences, and associations, could hinder a rapid and effective devel- The main goal of the meeting was to discuss the opment of AI, both research-wise and in real-life future of AI and brainstorm about possible devel- applications. opments, at several levels, to improve its positive impact on society. More precisely, the plan was AAAI (Association for the Advancement of AI) to discuss and concretely make steps forward on and IJCAI (International Joint Conference on AI) the following topics: are the two main associations worldwide that cover a very large spectrum of the AI topics. • How IJCAI, AAAI, and other AI-related confer- They have recently joined forces to try to turn ences and associations can have a more fruit- these multiple research directions into an oppor- ful and effective collaboration; tunity, by fostering fruitful cooperations and posi- • The impact of AI on society; tive synergies among all these actors in the • The relationships among the various AI sub- world-wide AI arena. disciplines; • The future of AI as a discipline; As a ﬁrst step in this direction, as presidents of • The deﬁnition and organization of innovative these two associations, (Francesca Rossi, IJCAI events related to AI. and Manuela Veloso, AAAI) we co-organized a meeting to brainstorm about the future of AI as a The discussion was organized around four main discipline, the cooperation among the various themes:

17 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014

• AI disciplines and practice: Vision and accom- plishments of AI disciplines and their varied nature, in different parts of the world, in research, in education, and in practice. • Associations, conferences, journals: Relation- ship among AI associations; Conferences vs. journals; Role of large AI conferences. • Strong AI, Integrated AI, Open AI, and Grand Challenges. • Societal impact: Outreach; Funding strategies; Ethical issues in AI.

Before the meeting, the participants were invited and SAT, focus on a speciﬁc AI discipline. Their to prioritize their interests on each theme. We cooperation and synergic activities are often left then divided the participants across the four to the initiative of single people and not organ- themes, such that each would participate in a ized in structured ways. The summit’s partici- panel of at least one theme. At the meeting, the pants felt the need to investigate the possible discussion of a theme was scheduled to include birth of a new AI organization whose role would the very brief presentation of a panel (of about be to facilitate interaction and cooperation 12 participants), followed by a few questions, among all disciplines of AI, through various initia- and then, most importantly the discussion con- tives such as joint events, an open information tinued in six round tables led by two panelists structure for AI knowledge to foster integration, and concluded with the summary presented the common activities to outreach the society at members of each table. large and to have a positive impact, as well as guidelines for behavior of AI researchers about The discussion was very intense and productive, ethical issues in AI. A committee has been it and led to several important deliberations and formed, and it is currently brainstorming about plans for concrete actions to advance AI and im- this in order to put forward a concrete proposal prove its positive impact on society. In particular, soon to the whole community. besides several speciﬁc suggestions for concrete actions, there was the feeling that AI needs to be More details about the NY AI Summit and the united again, while maintaining the identity of the complete list of its participants can be found at several disciplines. http://ijcai.org/NY-AI-Summit/ .

There are many AI-related organizations and We would like to thank all the Summit’s partici- disciplines. Some, like AAAI, IJCAI, SIGAI, and pants for their very enthusiastic response to our ECCAI, encompass most of the AI spectrum, invitation to participate, as well as the tremen- while others, like ACP, ICAPS, KR, ML, Robotics, dous energy and positive attitude during the

18 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014

Summit discussions. We would also like to of CP 2003 and of IJCAI 2013. She is in the thank the sponsors that generously helped in the editorial board of Constraints, Artiﬁcial Intelli- organization of this meeting: AAAI, IJCAI, the gence, JAIR, AMAI, and KAIS. She has pub- Artiﬁcial Intelligence Journal, Elsevier, and the lished more than 160 papers on international Center for Urban Science and Progress (CUSP) journals and conferences, she co-authored one at NYU, where the meeting took place. Helmut book, and she co-edited several volumes, Simonis is credited for the photos. among which the handbook of constraint programming.

Manuela Veloso is Herbert A. Simon University Professor at Carnegie Mellon University. She researches in artiﬁcial intelligence and robotics, in particular on agents that Collaborate, Ob- serve, Reason, Act, and Learn (CORAL group). She is Fellow of AAAI, IEEE, and AAAS. Veloso co-founded RoboCup, a worldwide initiative in- Francesca Rossi is professor of computer sci- vestigating teams of autonomous robots in highly ence at the University of Padova, Italy. Her re- uncertain environments. With her students, real- search interests include constraint reasoning, izing that autonomous robots inevitably have preferences, multi-agent systems, computational limitations in perception, cognition, and action, social choice, artiﬁcial intelligence. She is both Veloso introduced symbiotic autonomous robots an ECCAI and a AAAI fellow. She has been the that can proactively ask for help from humans, president of the international association for con- other AI agents, and the web. For additional in- straint programming (ACP) and she is now the formation, including publications, see president of IJCAI. She has been program chair www.cs.cmu.edu/~mmv .

19 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014

V Drop-in Games at RoboCup Peter Stone (Dept. of Computer Science, The University of Texas at Austin; [email protected]) Patrick MacAlpine ([email protected]) Katie Genter ([email protected]) Sam Barrett ([email protected]) DOI: 10.1145/2639475.2655756

This picture shows a snapshot from one of the domly selecting different RoboCup teams to ﬁrst “Drop-in Challenge” games that was held at each contribute one robot to a team that plays RoboCup 2013 in Eindhoven, The Netherlands. against another such team. The robots must be Typically, RoboCup soccer games involve a programmed to work with previously unknown team of robots programmed by one university teammates. against a team programmed by another. As such, the teamwork strategies can all be “pro- In this picture, each robot was programmed by a grammed in.” However, as robots and their different RoboCup team. At RoboCup 2014, in agents become more capable of long-term Joao de Pessoa Brazil, the drop-in challenge autonomy, there will be increasing opportunities was repeated as a formal competition (from July and need for “ad hoc teamwork” in which agents 20th to July 24th, 2014). Afterwards, the top ﬁve need to cooperate without prior coordination. players in the competition were put together on The drop-in challenge at RoboCup provides an an all-star team to play against the 2014 cham- opportunity to study ad hoc teamwork by ran- pions of the main (full team) competition.

20 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014

A paper documenting the 2013 drop-in challenge Patrick MacAlpine is a fifth will be presented at IROS 2014 in September: year Computer Science http://www.cs.utexas.edu/~pstone/Papers/bib2h Ph.D. student at the Univer- tml/b2hd-IROS14-MacAlpine.html sity of Texas at Austin. He A paper describing ad hoc teamwork as an AI received his B.S. and MSEE challenge was presented at AAAI 2010: degrees in Electrical Engi- http://www.cs.utexas.edu/~pstone/Papers/bib2h neering from Rice Univer- tml/b2hd-AAAI10-adhoc.html sity. He is a member of the Learning Agents Research Group (LARG), advised by Dr. Peter Stone is an Peter Stone. His current focus is on using rein- Alfred P. Sloan Research forcement learning to develop locomotion skills Fellow, Guggenheim Fel- and strategy for the the UT Austin Villa RoboCup low, AAAI Fellow, Ful- 3D Simulation League team. He is currently sup- bright Scholar, and Uni- ported by a NDSEG fellowship. versity Distinguished Teaching Professor in Katie Genter is a fifth year the Department of Com- Computer Science Ph.D. puter Science at the Uni- student working on multia- versity of Texas at Aus- gent systems research in tin. He received his Dr. Peter Stone’s Learning Ph.D. in Computer Sci- Agents Research Group at ence in 1998 from Car- the University of Texas at negie Mellon University. Austin. Her research fo- From 1999 to 2002 he was a Senior Technical cuses on how to design Staff Member in the Artificial Intelligence Princi- agents that can be added to a pre-existing team ples Research Department at AT&T Labs - Re- and influence that team to behave in a particular search. Peter's research interests include ma- way. She specifically studies adding controllable chine learning, multiagent systems, robotics, and agents to flocks, where the non-controllable e-commerce. In 2003, he won a CAREER award agents in the flock determine their orientation from the National Science Foundation for his re- based on the agents nearest to them. She works search on learning agents in dynamic, collabora- to design algorithms for the controllable agents tive, and adversarial multiagent environments. such that they can influence the flock to travel In 2004, he was named an ONR Young Investi- in a particular way, such as to avoid obstacles gator for his research on machine learning on during migration. Before beginning her Ph.D., physical robots. In 2007, he was awarded the she obtained a bachelors degree in Computer prestigious IJCAI 2007 Computers and Thought Science from the Georgia Institute of Technology award, given once every two years to the top AI in 2009. researcher under the age of 35. In 2013 he was awarded the University of Texas System Re- Samuel Barrett is a Ph.D. gents' Outstanding Teaching Award and in 2014 candidate at the University he was inducted into the UT Austin Academy of of Texas at Austin. He is a Distinguished Teachers. member of the Learning Agents Research Group (LARG) led by Peter Stone. In 2012, Sam led the UT Austin Villa team to win the international RoboCup championship for robot

21 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014 soccer in the Standard Platform League (SPL). teammates on the ﬂy. His interests also include In 2009, he received an NDSEG graduate fel- machine learning, multiagent systems, and ro- lowship. He received his B.S. in Computer Sci- botics. After graduating, Sam will join Kiva Sys- ence in 2008 from Stevens Institute of Technol- tems. ogy. His research focuses on ad hoc teamwork, enabling robots and other agents to adapt to new

22 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014

V Visualizing Information Networks: The 2011 Fukushima Earthquake Manuel Gomez Rodriguez (Max Planck Institute for Software Systems; [email protected]) Jure Leskovec (Stanford University; [email protected]) DOI: 10.1145/2639475.2655757

These networks visualize the way in which me-   mes about Fukushima's earthquake propagate  among blogs (in red) and mainstream media sites (in blue) at three different points in ti-  me (March 18, 2011; June 25, 2011; and October 13, 2011), as inferred by InfoPath. Infopath is a network inference algorithm that infers “who- copies-from-whom” from massive crawls of the Web. The inferred networks give insights  about the position and roles various sites play in the spread of ideas over the Web as well as helping us understand how information path- ways change over time. Infopath was developed March 18, 2011 by researchers at Stanford University and Max Planck Institute for Intellingent Systems. Learn more about it at http://snap.stanford.edu/infopath .

  M a n u e l G o m e z Rodriguez is a tenure- track research group leader at Max Planck In- stitute for Software Sys- tems. Manuel develops machine learning and   large-scale data mining  June 25, 2011 methods for the analysis  and modeling of large real-world networks and

 processes that take place   over them. He is particu-   larly interested in problems arising in the Web  and social media and has received several rec- ognitions for his research, including an Outstand- ing Paper Award at NIPS’13 and a Best Re-

 search Paper Honorable Mention at KDD'10.   Manuel holds a Ph.D. in Electrical Engineering from Stanford University and a B.S. in Electrical Engineering from Carlos III University in Madrid (Spain). You can ﬁnd out more about him at October 13, 2011 http://people.tuebingen.mpg.de/manuelgr/ .

23 AI MATTERS, VOLUME 1, ISSUE 1!AUGUST 2014

Jure Leskovec is assistant ulty Fellowship, the Alfred P. Sloan Fellowship, professor of Computer and best paper awards at KDD, WSDM, WWW, Science at Stanford Uni- and ICDM. Leskovec received his bachelor’s versity. His research fo- degree in computer science from the University cuses on mining large so- of Ljubljana, Slovenia, and his Ph.D. in in ma- cial and information net- chine learning from the Carnegie Mellon Univer- works. Problems he inves- sity and postdoctoral training at Cornell Univer- tigates are motivated by sity. You can follow him on Twitter @jure or visit large scale data, the Web http://cs.stanford.edu/~jure/ . and on-line media. This research has won several awards including a Microsoft Research Fac-