We finally have taught rats capitalism! And, fascinating, they behave a lot like humans in the stock market, displaying human investment biases including doubling down on losses & a tendency to avoid quickly realizing losses while rapidly realizing gains. https://
psycnet.apa.org/record/2021-15
335-002
…
RESEARCH
-
Rats Display Human Investment Biases in Capitalism Study
By
–
-
China’s NLP Innovation Challenge: Breaking Rules Beyond Incremental Improvements
By
–
They’ve been very active in NLP & Deep Reinforcement Learning too. They’ve been very good on taking existing techniques & improving them. The challenge will be to innovate (ie break the rules) – same challenge Japan had in early 90s + President Xi clamped down on tech innovators.
-
Simulation Environments with Thermodynamics for RL Research
By
–
Question: Can anyone recommend simulation environments for AI research where the second law of thermodynamics applies? It feels like this is needed to avoid collapse of some RL exploration strategies and intrinsic motivation. https://
flip.it/SodkpS -
AI protein structure predictions advance cancer research drug discovery
By
–
“3D protein structure predictions made by AI could boost cancer research and drug discovery” — let’s hope https://
flip.it/357CH1 -
Prompt-to-Prompt Brings Image Editing to Text-to-Image Models
By
–
Prompt-to-Prompt — “Brings Image Editing Capabilities to Text-to-Image Models” https://
flip.it/VzV4OT -
Formal Informal Languages: Modern Prompt Engineering Approaches
By
–
Formal Informal Languages — An interesting take on (current) prompt engineering https://
flip.it/lmLQod -
Inductive Biases Deep Learning Higher Level Cognition Overview
By
–
“Inductive biases for deep learning of higher-level cognition” by Anirudh Goyal and Yoshua Bengio has a good overview of many important ideas in deep learning and the study of intelligence. https://
royalsocietypublishing.org/doi/pdf/10.109
8/rspa.2021.0068
… -
OpenReview Verification Launch Without Cost
By
–
Just noticed this green check on @openreviewnet
, in the Author Console OpenReview verified baby, and it cost $0 Your move, @elonmusk
. -
Cerebras Wafer-Scale Cluster Achieves Linear Scaling Performance
By
–
Distributed training across thousands of traditional accelerators suffers diminishing returns as more compute is added. Join Natalia Vassilieva at #SC22 as she presents the Cerebras Wafer-Scale Cluster, which achieves near-perfect linear scaling across additional cores
#AI #SC22 -
Big Data Big Responsibility: Ethics in AI and Machine Learning
By
–
as is often said – with great power comes great responsibility ! Big Data, Big Responsibility – DATAVERSITY https://
buff.ly/3NOvgiW
#bigdata #ai #ArtificialIntelligence #MachineLearning #deeplearning #data