As researchers across @UUtah build, use, & study AI, they’re uncovering high-stakes ethical questions that can’t be solved by technologists or humanists alone. They recently worked through such Qs together at a SCI-hosted workshop: https://t.co/6qotJmFVBF
Calling @UUtah researchers across tech, humanities, & social sciences: we’ve expanded capacity for our April 3 AI + Ethics Workshop. Discover AI work beyond your field & shape new approaches to some of the most important ethical challenges of our time: https://t.co/OZFBK3dwDt
Please join us for today's at 11 AM MST Data Science and AI lecture series to hear Varun Shankar in his talk titled "Kernel Methods for Operator Learning." We look forward to seeing you either in person or online! https://t.co/R9HbBpuM26 #datascience#ai#ucdsai#ml
Jeff M. Phillips @KahlertSoC:
Mathematical Foundations for Data Analysis (2021)
#SpringerSeriesintheDataScience#DataScience#CGWeek23
🔓Free view 🔗Chapter 6. Gradient Descent https://t.co/ZX4OW099zx
Textbook, suitable for an early undergraduate up to a graduate course 🌿
If you are in Leipzig, Germany, next week, please check out my talk (https://t.co/Abvco1cbmB), on Ferret: Reviewing Tabular Datasets for Manipulation! H/T Shaurya Sahai, @probablyjeff, and @alexander_lex. 🧵 (1/8)
Data science is amongst one of the most interdisciplinary areas of the research sciences. Explore a selection of featured data science books and journals on the topics of business, economics & finance for free through January 15th! https://t.co/6Chyu3wOFR
Congratulations to @UUtah's @keviv9 (from @UtahSoC & @UUDataScience) on being named one of the 2021-2022 @Bloomberg#DataScience Ph.D. Fellows!
Learn more about his research focus and the other Fellows in this year's cohort: https://t.co/xOzI5aAwaK
#AI#ML#NLProc
It's time to stop making t-SNE & UMAP plots. In a new preprint w/ Tara Chari we show that while they display some correlation with the underlying high-dimension data, they don't preserve local or global structure & are misleading. They're also arbitrary.🧵https://t.co/XkAOTKlOcs
3/3 It’s a balanced approach of bias mitigation and information retention. Our intrinsic and extrinsic evaluations on gender-occupation bias show it performs on par with different projective methods while retaining more coherent information.
2/3 Projective debiasing can be too aggressive as it removes extra information that should actually be preserved. To address this, we propose a graded rotation operation to correct and rectify biases in embedding space (GloVe/RoBERTa).