Ben WaberI had a lot of time waiting around for my oldest during activities today, which meant some long walks with great talks for my #AcademicRunPlaylist! (1/12)
First was an interesting talk by Hao Su on modeling the physical world for embodied #AI. The work presented here is representative of a recent intriguing trend in #robotics - attempting to translate the success of large datasets in non-embodied AI to robots, with some impressive results shown here https://www.youtube.com/watch?v=nkTdQ3oFZ0Q&t=28s (2/12)
Winter 2023 Robotics Colloquium: Hao Su (UCSD)
Next was a detailed set of talks on the socio-economic impacts of COVID-19 in #Cambodia at the #Crawford School of Public Policy. Overall, it seems that Cambodia saw similar effects to other countries in southeast Asia, with the caveat that geographic unemployment patterns seem highly uneven https://www.youtube.com/watch?v=hkhzzDLZO5I (3/12) #economics
FKP 2022 06 30 - Socio-economic impacts of COVID-19 in Cambodia: some snapshots
Next was a nice slate of talks on rethinking the protections of vulnerable #workers at #FondazioneMarcoBiagi. I particularly liked Camilla Spadavecchia's talk on the impact of COVID-19 on high skilled migrants in the #Netherlands https://www.youtube.com/watch?v=MSu0xBZZneI (4/12)
Rethinking the Protections of Vulnerable Workers
Next was a technical talk by Chandra Thapa on enabling #DeepLearning with #FederatedLearning and split learning at #Monash Information Technology. I still wonder about #privacy protection mechanisms like the one presented here accelerating discrimination by making everyone look more like the majority class, but the architecture itself is quite interesting https://www.youtube.com/watch?v=RYuWpG41YmU (5/12) #AI
Deep Learning in a Resource-Limited Environment with Privacy and Security | Cybersecurity Seminars
Next was a fascinating talk by Ioannis Kaloskampis on using shipping #data for real time economic activity analysis at the #RoyalStatisticalSociety. I suspect a lot of folks in investment banking are already doing this, but the detailed description of how to calculate these indicators was incredible, and indicates how alternative metrics can provide more timely, accurate metrics than traditional econ stats. Highly recommend https://www.youtube.com/watch?v=OwqDDwsjnpM (6/12) #economics
Faster indicators of UK economic activity: Shipping
David RobinsonNext was an excellent panel on #AI #bias prevention at the #RhodesTrust with Elaine Nsoesie, @dgrobinson, and Muhammed Razzak. The topics covered here, from equity to algorithm design to prioritizing particular bias problems, are expansive and treated with the care they deserve. Highly recommend https://www.youtube.com/watch?v=yH76s6BIi7c (7/12) #AIEthics
Tech & Society 2022: 6A. Equity: Bias Prevention, not just Mitigation
Next was a feisty panel on the role of #antitrust in mediating #tech company power at the #SiliconFlatirons w/Kristelia García, @elibietti, Renata B. Hesse, @halsinger, and Christopher Yoo. I would've liked more discussion of alternatives to the consumer welfare standard, still and interesting conversation overall https://www.youtube.com/watch?v=M1CA3P0VSFU (8/12)
Next was an engaging discussion with Eric Beinhocker and @DianeCoyle1859 on rethinking #economics for the modern world at the @sfiscience. The framing here, which clearly articulates the problem of the power of modern monopolies and what metrics to focus on, is essential. Highly recommend https://www.youtube.com/watch?v=rB6CkhOglLI&t=1s (9/12)
Eric Beinhocker & Diane Coyle on Rethinking Economics for A Sustainable & Prosperous World (EPE 02)
Next was a short talk by Agnieszka Kanas on soft #skills and immigrants' economic outcomes at @GESIS. Soft skills seem to significantly improve economic outcomes for immigrants, but less so for nonwestern #immigrants (likely due to bias) https://www.youtube.com/watch?v=mN3VwfvvgJU (10/12)
Agnieszka Kanas: Soft skills and immigrants’ economic outcomes in Europe
Next was a fantastic talk by @dan on building tools for data-driven #labor organizing. Algorithmic #management isn't just scientifically a terrible idea, it inflicts real harm on #workers who often have no recourse. Calacci's impressive work and collaboration with organizers shows that worker-collected data can begin to rebalance this equation and provide more insight into opaque algorithmic management systems. Highly recommend https://www.youtube.com/watch?v=z8rJN4ikQhQ (11/12) #PeopleAnalytics
A New HOPE (2022): How to Bargain With a Black Box: Hacking a Path to Data-Driven Organizing
Last was a timely panel on human and machine #design with Amelie Klein, Gilles Retsin, and Ariana Dongus at the #Vitra Design Museum. This conversation considers the implications of generative #AI from a variety of perspectives that seems lacking in tech circles - designers, gig workers, #art, and society more broadly. Highly recommend https://www.youtube.com/watch?v=YWT-6wwqMLs (12/12) #AIEthics #AIArt #GenerativeArt #GenerativeAI
