Next Meeting - Tuesday, October 10th - Adam Russell "Deep Learning in Perl"
LIVE Speaker ! ROOM Change (from spring): E51-372 "Deep Learning with AI::MXNet: Navigating implementation issues" ABSTRACT This talk will cover lessons learned from a recent experience in getting a deep learning projected started, with little prior experience in AI. All code will use the Perl MXNet API and guide the audience through developing a simple model, which is then built on to perform more complex tasks. Artificial Intelligence (AI) has for a long time captured the popular imagination. Results from academia and industry have finally started to deliver on some of the long hoped for results: self driving cars, automated medical diagnoses, and written and verbal language processing. These areas are showing advances that were once simply products of fiction writers. The current wave of AI enthusiasm may be attributed to what is called Deep Learning which is a convenient label for relatively new techniques using neural networks. The possibility for increased automation across virtually every industry has resulted in the spinning up of many new startup companies, as well as new projects within existing enterprises, resulting in the need to develop the skills necessary to pursue this new area. While not the language of production, Perl is used to develop algorithms and demonstrate concepts before they receive fuller treatment. Deep Learning practitioners often begin their deep learning work, correctly, with a review of the literature and research into the fundamentals. Projects then often start confidently with high hopes, built on that conceptual understanding, only to quickly get bogged down in unforeseen, but critically important, issues of implementation. SPEAKER Adam Russell is a software engineer with OptumLabs' Center for Applied Data Science (CADS). CADS is tasked with developing prototype applications which implement recent advances in algorithms and technology to address issues of importance to Optum business interests. Most recent projects have been focused on Deep Learning. Adam has a PhD in Computer Science from the University of Massachusetts Lowell, his academic interests involve Computational Geometry and Data Visualization and these explorations, much like the work described in this talk, are all Perl driven. He also teaches, on an adjunct basis, at Wentworth Institute of Technology in Boston. NOTES Parking Alert. Recent changes in MIT Parking Dept web pages (parking , visitors, public ) no longer allow un-permitted parking after-hours. (This is a natural response to several other East Campus parking lots being eaten by new building sites for campus expansion.) Only legit parking is Cambridge meters (late hours! and several blocks lost to construction), free spaces on Memorial Drive (allow time for circling), and paid lot/garages. Come by Train, Bus, Bicycle, or Foot if you can! (Parking at MBTA Garages are convenient to T and not overpriced, unlike most in-town garages.) Also, construction detour is even longer than last time if driving, Wadsworth to/from Mem Drive & Amherst St to Amess both closed, only access by car is Wadsworth to/from Main St east-bound. DETAILS Tech Meetings are held on the 2nd Tuesday of every month at MIT building E51, Sloan School Tang Center [not the other Tang building!] nearer to Kendall Sq than Mass Ave. (directions). Talk begins at 7:30. Refreshments in the hallway prior. RSVP for count encouraged but not required, to bill.n1...@gmail.com or Boston-PM list, by 4pm Tuesday. (NOTE: Fall 2017: we're moving back to the squarer room 372 (first door after the partition), not the wider 376 (second door) that we had the last several years) Future - Fall/Winter reservations *If you have a demo or talk idea, please, when would you like to present? Doesn't need to fill the full time.* Confirmed room assignments courtesy of MIT Room E51-372 Tues, Nov 14th Tues, Dec 12th , 2017 _______________________________________________ Boston-pm mailing list Boston-pm@mail.pm.org http://mail.pm.org/mailman/listinfo/boston-pm
