Medijski sadržaj
- Tweetovi
- Tweetovi i odgovori
- Medijski sadržaj, trenutna stranica.
-
Dylan
@dylanslack20 starting his talk at#FAT2020 ! Paper: https://arxiv.org/abs/1908.09092 pic.twitter.com/BGwpIsgal7
-
The Tech & Justice series
@haverfordedu this semester starts NEXT WEEK with a keynote by Shoshana Zuboff@shoshanazuboff ! Public Talk: Thurs Feb 6 at 7pm The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power Map: https://goo.gl/maps/BDeX9naohR2JPueH8 …pic.twitter.com/QOfTNuSWuS
Prikaži ovu nit -
Presenting our work on measuring the environmental impact of code
@ClimateChangeAI workshop#NeurIPS2019 with@Medullosa today! Stop by East ballroom C. https://twitter.com/kdphd/status/1197166057372471296 …pic.twitter.com/o0sU6C8iZL
-
Really nice paper presented today at the HCML
#NeurIPS2019 workshop by @CharlieMarx9@berkustun Terrifying results show large percentage of people can receive different recidivism prediction results from models that have essentially the same accuracy. https://arxiv.org/abs/1909.06677 pic.twitter.com/RwDJguf4l6
-
.
@dylanslack20 with his THREE posters / papers at the HCML#NeurIPS2019 workshop today! https://arxiv.org/abs/1902.03501 https://arxiv.org/abs/1908.09092 pic.twitter.com/Js2iYyGc11
-
Really nice to see all the interest in talking to @CharlieMarx9 and
@rlanasphillips about our work on indirect feature influence today at#NeurIPS2019 ! Paper: https://arxiv.org/abs/1906.08652 https://twitter.com/dylanslack20/status/1205286379204145152 …pic.twitter.com/4xi12UMOma
-
Code takes energy to run and long-running jobs have large CO2 footprints. How can you easily measure and communicate the energy usage and CO2 emissions of your code? New paper
@climatechangeai workshop, Python package, and associated curriculum! https://arxiv.org/abs/1911.08354 pic.twitter.com/PqHOoLHQnB
Prikaži ovu nit -
Tomorrow is the last event for the fall part of the Tech & Justice series
@haverfordedu:@thistimeitsmimi will be on campus to talk about her work In Absentia: Mimi Onuoha. Hope to see you there! Exhibit up through Nov 22. ...and stay tuned for more of the series next semester! https://twitter.com/kdphd/status/1171844490883194884 …pic.twitter.com/XC7WpCyjxu
Prikaži ovu nit -
And we'll close the semester with an exhibit by
@thistimeitsmimi Mimi Onuoha presents In Absentiapic.twitter.com/DGXe4KjUqC
Prikaži ovu nit -
The day after discussing Criminal In/Justice and Data Use, on Oct 25th, we'll have a workshop to organize for community power with
@hannahsassaman Fighting RATs: Organizing for Community Power over Risk Assessment Tools in Criminal Justicepic.twitter.com/wKv8x50SBP
Prikaži ovu nit -
We'll have a roundtable Oct 24th discussing Criminal In/Justice and Data Use with
@ruha9@KLdivergence@vmsoutherlandpic.twitter.com/jogHCT0ZYc
Prikaži ovu nit -
Keynote by
@zephoria will launch the series in just about two weeks from now! Thurs, Sept 26, 4:30pm danah boyd Vulnerabilities: How Social Media and Data Infrastructure are Exploited for Fun, Profit, and Politicspic.twitter.com/7CgTdung3R
Prikaži ovu nit -
This fall we'll be welcoming a bunch of amazing people to present
@haverfordedu on the theme of Tech & Justice! All events are open to the public. Come see@zephoria@ruha9@KLdivergence@vmsoutherland@hannahsassaman@thistimeitsmimi ! http://hav.to/techjustice pic.twitter.com/MPnthMfawU
Prikaži ovu nit -
Indirect influence handles feature importance in the face of correlated variables, like determining that race is important if a model uses zip code. Our technique takes a reductions approach, so any direct influence technique can be used. Code: https://github.com/charliemarx/disentangling-influence/ …pic.twitter.com/580k4LvGk4
Prikaži ovu nit -
Super proud that
@haverfordedu student @CharlieMarx9 had a paper accepted to#NeurIPS2019! Disentangling Influence applies disentangled reps to indirect influence of features on a model. Joint w/ alum@rlanasphillips, me,@scheidegger, and@geomblog. https://arxiv.org/pdf/1906.08652.pdf …pic.twitter.com/CMjGIFutWX
Prikaži ovu nit -
New paper: Automated Congressional Redistricting w/ former student Harry Levin! Tricky thing about automated redistricting: population deviation between real districts is often just 1 person (based on census data). We match this algorithmically.
#fairmaps http://sorelle.friedler.net/papers/redistricting.pdf …pic.twitter.com/IzCXRChnn7
Prikaži ovu nit -
Amazing first day at
#FAT2018 with current students and@haverfordedu alums!pic.twitter.com/yCidNJKtl7
-
Early predictive hiring for software engineers found spurious patterns that caused long-lasting discriminatory effects. We haven't yet moved past these issues - what harder-to-expose effects will the current automated hiring boom cause? https://www.bloomberg.com/news/features/2018-02-01/women-once-ruled-computers-when-did-the-valley-become-brotopia …pic.twitter.com/v0BivEefnB
-
Reminder: FAT* papers due tonight at midnight AoE! http://fatconference.org pic.twitter.com/KllZeiP2FV
-
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.