Goto

Collaborating Authors

Results


How to stop AI from recognizing your face in selfies

MIT Technology Review

A number of AI researchers are pushing back and developing ways to make sure AIs can't learn from personal data. Two of the latest are being presented this week at ICLR, a leading AI conference. "I don't like people taking things from me that they're not supposed to have," says Emily Wenger at the University of Chicago, who developed one of the first tools to do this, called Fawkes, with her colleagues last summer: "I guess a lot of us had a similar idea at the same time." Actions like deleting data that companies have on you, or deliberating polluting data sets with fake examples, can make it harder for companies to train accurate machine-learning models. But these efforts typically require collective action, with hundreds or thousands of people participating, to make an impact.


Artificial Intelligence: A Guide for Thinking Humans: Mitchell, Melanie: 9780374257835: Amazon.com: Books

#artificialintelligence

"Mitchell knows what she's talking about. Artificial Intelligence has significantly improved my knowledge when it comes to automation technology, [but] the greater benefit is that it has also enhanced my appreciation for the complexity and ineffability of human cognition."―John Warner, Chicago Tribune "Without shying away from technical details, this survey provides an accessible course in neural networks, computer vision, and natural-language processing, and asks whether the quest to produce an abstracted, general intelligence is worrisome . . . Mitchell's view is a reassuring one." AI isn't for the faint of heart, and neither is this book for nonscientists . . .


How machine learning can improve money management

#artificialintelligence

Two disciplines familiar to econometricians, factor analysis of equities returns and machine learning, have grown up alongside each other. Used in tandem, these fields of study can build effective investment-management tools, according to City University of Hong Kong's Guanhao Feng (a graduate of Chicago Booth's PhD Program), Booth's Nicholas Polson, and Booth PhD candidate Jianeng Xu. The researchers set out to determine whether they could create a deep-learning model to automate the management of a portfolio built on buying stocks that are expected to rise and short selling those that are expected to fall, known as a long-short strategy. They created a machine-learning algorithm that built a long-short equity portfolio from the top and bottom 20 percent of a 3,000-stock universe. They ranked the equities using the five-factor model of Chicago Booth's Eugene F. Fama and Dartmouth's Kenneth R. French.


Quarterly Workshop: Machine Learning and Strategic Behavior

#artificialintelligence

The Quarterly CS Econ Workshop brings in three or four experts at the interface between computer science and economics to present their perspective and research on a common theme. Chicago area researchers with interest in economics and computer science are invited to attend. The technical program is in the morning and includes coffee and lunch (on your own). The afternoon of the workshop will allow for continued discussion between attendees and the speakers. The workshop series is organized by Jason Hartline, Benjamin Golub, Annie Liang, Marciano Siniscalchi, and Alireza Tahbaz-Salehi.


Behavioral convergence in humans and animals

Science

Over the 20th century, the social sciences developed without taking much notice of humans' nature as products of evolution. In the 1970s this attitude was challenged by behavioral biologists ([ 1 ][1], [ 2 ][2]) who asserted that general principles concerning the behavior of life forms must also be relevant to understanding human behavior. They argued that because human cognition and emotions had evolved by natural selection, these behavior-generating mechanisms should generally shape behavior so that it maximizes biological fitness. Not all social scientists agreed. Cultural anthropologists, in particular, were mostly aghast at the rigidly scientific and overtly biological nature of this perspective, viewing it as blatantly flawed ([ 3 ][3]). They claimed that differences between and within human societies were mainly due to variant cultural belief systems. On page 292 of this issue, Barsbai et al. ([ 4 ][4]) show that adaptation to local ecological conditions is an important determinant of variation in human behavior in traditional societies. The sample analyzed by Barsbai et al. consists of 339 hunter-gatherer societies that are most appropriate for comparison because their members' lives and livelihoods are intimately constrained by the natural world. The authors show that variation in hunter-gatherer patterns for 15 behavioral variables statistically converges on the same characteristics that are most common in birds and mammals in the same local regions of the world. These traits include diet composition, mobility patterns, paternal investment, divorce rates, social group size, and social stratification. In other words, in places where hunter-gatherers are more polygynous, there also tend to be more polygynous bird and mammal species. These patterns appear to be driven by ecological and habitat similarity, not by locational proximity per se. Not only are hunter-gatherers behaviorally similar in similar ecologies, but even mammals and birds in those ecologies tend to exhibit the same behavioral regularities as do the human populations. Hence, the study appears to validate the basic premise of the evolutionary perspective called “human behavioral ecology” ([ 5 ][5], [ 6 ][6]). However, it is a mistake to conclude from this that culture is unimportant. Beginning in the 1980s, researchers in human behavioral sciences developed a sophisticated, scientific, evolutionary theory accounting for the role of culture in human behavior ([ 7 ][7], [ 8 ][8]). These scientists provided both theoretical and empirical evidence that social learning was a prime determinant of human behavioral variation. The affective and cognitive mechanisms that underpin social learning are adaptations and are in large part responsible for our species' spectacular ecological success, but they also create historical patterns absent in other species and lead to outcomes not predicted by theories developed for noncultural creatures. Barsbai et al. show convincingly that ecological factors explain much variation in human behavior, but so too does cultural history. For example, Mathew and Perreault ([ 9 ][9]) studied the causes of variation among 172 native American groups in western North America. Like Barsbai et al. , they found that ecological factors explained a substantial amount of variation, particularly in behaviors related to subsistence and technology. But the variation in subsistence-related behaviors was equally well explained by the linguistic distance between groups, which proved to be an even better explanation than ecological factors for the variation in political organization, religious practice, and kinship organization. Moreover, the effect of cultural history seems to persist for hundreds or even thousands of years. Cultural evolution can also lead to outcomes not predicted by the evolutionary mechanisms applied to other species. Barsbai et al. show that variation in human residential group size has the same relationship to ecology as in other species. However, human foragers are much more cooperative than other primates ([ 10 ][10]), and sometimes they cooperate in groups numbering hundreds of individuals in communal foraging, construction of shared capital facilities, and warfare ([ 11 ][11]). No other vertebrate cooperates on these scales. Exactly why this is the case is controversial, but it seems likely that culturally transmitted social norms play an important role. Culture and genes are linked in a tight coevolutionary embrace, and this leads to complex patterns of genetic and cultural co adaptation. For example, Henrich has recently argued ([ 12 ][12]) that the extent to which people are embedded in networks of kin obligation is a function of both ecological factors (cooperative intensive agriculture) and cultural history (church edicts against kin marriage and collective property institutions), and that variations in the intensity of kin-network embeddedness ultimately transformed human psychology into that observed today in “WEIRD” (Western, educated, industrialized, rich, democratic) societies, where individualism is paramount, rather than the psychology of traditional societies, where collectivism and kin-group favoritism predominate. Experiments show that the cognitive, emotional, and psychological effects of these different cultural histories are profound, and imply that findings from Western modern societies may often be irrelevant to predicting behavior in non-Western and traditional societies. Likewise, the spread of monogamy in modern societies, despite increasing wealth stratification, appears to be a puzzle that requires both adaptive modeling ([ 13 ][13]) and a recognition that monogamous social norms substantially increased cooperation with societies, and this norm can spread by group competition and cultural imitation ([ 12 ][12]). Coevolution between genes and culture in different ecologies may lead to uniquely human patterns not anticipated by animal studies. These examples illustrate just how complex human behavioral studies will become when the social sciences fully integrate an adaptive evolutionary view with a view of human behavioral variation in terms of cultural social norms. So far, a complete theory that predicts when culture will override fitness maximizing ecological adaptation and vice versa is not available. That will be the challenge for the next generation of social scientists, as an “either/or” view is replaced with an integrated evolutionary theory of human behavior. 1. [↵][14]1. E. O. Wilson , Sociobiology (Harvard Univ. Press, 1976). 2. [↵][15]1. R. Alexander , Darwinism and Human Affairs (Univ. of Washington Press, 1979). 3. [↵][16]1. M. Sahlins , The Use and Abuse of Biology (Univ. of Michigan Press, 1976). 4. [↵][17]1. T. Barsbai, 2. D. Lukas, 3. A. Pondorfer , Science 371, 292 (2021). [OpenUrl][18][Abstract/FREE Full Text][19] 5. [↵][20]1. B. Winterhalder, 2. E. Smith , Evol. Anthropol. 9, 51 (2000). [OpenUrl][21][CrossRef][22][Web of Science][23] 6. [↵][24]1. M. Borgerhoff Mulder , Behav. Ecol. 24, 1042 (2013). [OpenUrl][25][CrossRef][26] 7. [↵][27]1. R. Boyd, 2. P. Richerson , Culture and the Evolutionary Process (Univ. of Chicago Press, 1985). 8. [↵][28]1. L. Cavalli-Sforza, 2. M. Feldman , Cultural Transmission and Evolution (Princeton Univ. Press, 1981). 9. [↵][29]1. S. Mathew, 2. C. Perreault , Proc. R. Soc. B 282, 20150061 (2015). [OpenUrl][30][CrossRef][31][PubMed][32] 10. [↵][33]1. K. Hill , Hum. Nat. 13, 105 (2002). [OpenUrl][34][CrossRef][35][Web of Science][36] 11. [↵][37]1. R. Boyd , A Different Kind of Animal (Princeton Univ. Press, 2018). 12. [↵][38]1. J. Henrich , The WEIRDest People in the World (Farrar, Straus and Giroux, 2020). 13. [↵][39]1. C. Ross et al ., J. R. Soc. Interface 15, 144 (2018). [OpenUrl][40] [1]: #ref-1 [2]: #ref-2 [3]: #ref-3 [4]: #ref-4 [5]: #ref-5 [6]: #ref-6 [7]: #ref-7 [8]: #ref-8 [9]: #ref-9 [10]: #ref-10 [11]: #ref-11 [12]: #ref-12 [13]: #ref-13 [14]: #xref-ref-1-1 "View reference 1 in text" [15]: #xref-ref-2-1 "View reference 2 in text" [16]: #xref-ref-3-1 "View reference 3 in text" [17]: #xref-ref-4-1 "View reference 4 in text" [18]: {openurl}?query=rft.jtitle%253DScience%26rft.stitle%253DScience%26rft.aulast%253DBarsbai%26rft.auinit1%253DT.%26rft.volume%253D371%26rft.issue%253D6526%26rft.spage%253D292%26rft.epage%253D295%26rft.atitle%253DLocal%2Bconvergence%2Bof%2Bbehavior%2Bacross%2Bspecies%26rft_id%253Dinfo%253Adoi%252F10.1126%252Fscience.abb7481%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [19]: /lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6Mzoic2NpIjtzOjU6InJlc2lkIjtzOjEyOiIzNzEvNjUyNi8yOTIiO3M6NDoiYXRvbSI7czoyMjoiL3NjaS8zNzEvNjUyNi8yMzUuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9 [20]: #xref-ref-5-1 "View reference 5 in text" [21]: {openurl}?query=rft.jtitle%253DEvol.%2BAnthropol.%26rft.volume%253D9%26rft.spage%253D51%26rft_id%253Dinfo%253Adoi%252F10.1002%252F%2528SICI%25291520-6505%25282000%25299%253A2%253C51%253A%253AAID-EVAN1%253E3.0.CO%253B2-7%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [22]: /lookup/external-ref?access_num=10.1002/(SICI)1520-6505(2000)9:2<51::AID-EVAN1>3.0.CO;2-7&link_type=DOI [23]: /lookup/external-ref?access_num=000086885300001&link_type=ISI [24]: #xref-ref-6-1 "View reference 6 in text" [25]: {openurl}?query=rft.jtitle%253DBehav.%2BEcol.%26rft_id%253Dinfo%253Adoi%252F10.1093%252Fbeheco%252Fars223%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [26]: /lookup/external-ref?access_num=10.1093/beheco/ars223&link_type=DOI [27]: #xref-ref-7-1 "View reference 7 in text" [28]: #xref-ref-8-1 "View reference 8 in text" [29]: #xref-ref-9-1 "View reference 9 in text" [30]: {openurl}?query=rft.jtitle%253DProc.%2BR.%2BSoc.%2BB%26rft_id%253Dinfo%253Adoi%252F10.1098%252Frspb.2015.0061%26rft_id%253Dinfo%253Apmid%252F26085589%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [31]: /lookup/external-ref?access_num=10.1098/rspb.2015.0061&link_type=DOI [32]: /lookup/external-ref?access_num=26085589&link_type=MED&atom=%2Fsci%2F371%2F6526%2F235.atom [33]: #xref-ref-10-1 "View reference 10 in text" [34]: {openurl}?query=rft.jtitle%253DHum.%2BNat.%26rft.volume%253D13%26rft.spage%253D105%26rft_id%253Dinfo%253Adoi%252F10.1007%252Fs12110-002-1016-3%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [35]: /lookup/external-ref?access_num=10.1007/s12110-002-1016-3&link_type=DOI [36]: /lookup/external-ref?access_num=000176497800005&link_type=ISI [37]: #xref-ref-11-1 "View reference 11 in text" [38]: #xref-ref-12-1 "View reference 12 in text" [39]: #xref-ref-13-1 "View reference 13 in text" [40]: {openurl}?query=rft.jtitle%253DJ.%2BR.%2BSoc.%2BInterface%26rft.volume%253D15%26rft.spage%253D144%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx


Infrastructure for Artificial Intelligence, Quantum and High Performance Computing

arXiv.org Artificial Intelligence

William Gropp (University of Illinois at Urbana-Champaign), Sujata Banerjee (VMware Research) and Ian Foster (University of Chicago) High Performance Computing (HPC), Artificial Intelligence (AI)/Machine Learning (ML), and Quantum Computing (QC) and communications offer immense opportunities for innovation and impact on society. Researchers in these areas depend on access to computing infrastructure, but these resources are in short supply and are typically siloed in support of their research communities, making it more difficult to pursue convergent and interdisciplinary research. Such research increasingly depends on complex workflows that require different resources for each stage. This paper argues that a more-holistic approach to computing infrastructure, one that recognizes both the convergence of some capabilities and the complementary capabilities from new computing approaches, be it commercial cloud to Quantum Computing, is needed to support computer science research. The types of infrastructure needed to support HPC and AI/ML share many features; GPU systems originally developed for HPC have become essential for ML, and those systems have further been optimized for ML, with features now being applied to HPC simulations.


GAEA: Graph Augmentation for Equitable Access via Reinforcement Learning

arXiv.org Artificial Intelligence

Disparate access to resources by different subpopulations is a prevalent issue in societal and sociotechnical networks. For example, urban infrastructure networks may enable certain racial groups to more easily access resources such as high-quality schools, grocery stores, and polling places. Similarly, social networks within universities and organizations may enable certain groups to more easily access people with valuable information or influence. Here we introduce a new class of problems, Graph Augmentation for Equitable Access (GAEA), to enhance equity in networked systems by editing graph edges under budget constraints. We prove such problems are NP-hard, and cannot be approximated within a factor of $(1-\tfrac{1}{3e})$. We develop a principled, sample- and time- efficient Markov Reward Process (MRP)-based mechanism design framework for GAEA. Our algorithm outperforms baselines on a diverse set of synthetic graphs. We further demonstrate the method on real-world networks, by merging public census, school, and transportation datasets for the city of Chicago and applying our algorithm to find human-interpretable edits to the bus network that enhance equitable access to high-quality schools across racial groups. Further experiments on Facebook networks of universities yield sets of new social connections that would increase equitable access to certain attributed nodes across gender groups.


Medical imaging, AI, and the cloud: what's next? - Microsoft Industry Blogs

#artificialintelligence

Today marks the start of RSNA 2020, the annual meeting of the Radiological Society of North America. I participated in my first RSNA 35 years ago and I am super excited--as I am every year--to reconnect with my radiology colleagues and friends and learn about the latest medical and scientific advances in our field. Of course, RSNA will be very different this year. Instead of traveling to Chicago to attend sessions and presentations, and wander the exhibits, I'll experience it all online. While I will miss the fun, excitement, and opportunities to connect that come with being there in person, I am amazed by what a rich and comprehensive conference the organizers of RSNA 2020 have put together using the advanced digital tools that we have at hand now.


Holt-Winters Forecasting for Dummies (or Developers) - Part I - Gregory Trubetskoy

#artificialintelligence

This three part write up [Part II Part III] is my attempt at a down-to-earth explanation (and Python code) of the Holt-Winters method for those of us who while hypothetically might be quite good at math, still try to avoid it at every opportunity. I had to dive into this subject while tinkering on tgres (which features a Golang implementation). And having found it somewhat complex (and yet so brilliantly simple), figured that it'd be good to share this knowledge, and in the process, to hopefully solidify it in my head as well. Triple Exponential Smoothing, also known as the Holt-Winters method, is one of the many methods or algorithms that can be used to forecast data points in a series, provided that the series is "seasonal", i.e. repetitive over some period. In 1957 an MIT and University of Chicago graduate, professor Charles C Holt (1921-2010) was working at CMU (then known as CIT) on forecasting trends in production, inventories and labor force.


The Future of Artificial Intelligence

#artificialintelligence

June 8, 2019 Updated: April 20, 2020 "[AI] is going to change the world more than anything in the history of mankind. AI oracle and venture capitalist Dr. Kai-Fu Lee, 2018 In a nondescript building close to downtown Chicago, Marc Gyongyosi and the small but growing crew of IFM / Onetrack.AI have one rule that rules them all: think simple. The words are written in simple font on a simple sheet of paper that's stuck to a rear upstairs wall of their industrial two-story workspace. Sitting at his cluttered desk, located near an oft-used ping-pong table and prototypes of drones from his college days suspended overhead, Gyongyosi punches some keys on a laptop to pull up grainy video footage of a forklift driver operating his vehicle in a warehouse. It was captured from overhead courtesy of a Onetrack.AI "forklift vision system." The Future of Artificial Intelligence Artificial intelligence is impacting the future of virtually every industry and every human being. Artificial intelligence has acted as the main driver of emerging technologies like big data, robotics and IoT, and it will continue to act as a technological innovator for the foreseeable future. Employing machine learning and computer vision for detection and classification of various "safety events," the shoebox-sized device doesn't see all, but it sees plenty. Like which way the driver is looking as he operates the vehicle, how fast he's driving, where he's driving, locations of the people around him and how other forklift operators are maneuvering their vehicles. IFM's software automatically detects safety violations (for example, cell phone use) and notifies warehouse managers so they can take immediate action. The main goals are to prevent accidents and increase efficiency. The mere knowledge that one of IFM's devices is watching, Gyongyosi claims, has had "a huge effect." Marc Gyongyosi Photo Credit: IFM/OneTrack.AI The lower level of IFM was designed to mimic a warehouse environment so products can be effectively tested on site. Photo Credit: IFM/OneTrack.AI "If you think about a camera, it really is the richest sensor available to us today at a very interesting price point," he says. "Because of smartphones, camera and image sensors have become incredibly inexpensive, yet we capture a lot of information.