Befuddling AI Go Systems: MIT, UC Berkeley & FAR AI’s Adversarial Policy Achieves a >99% Win Rate Against KataGo https://
syncedreview.com/2022/11/02/bef
uddling-ai-go-systems-mit-uc-berkeley-far-ais-adversarial-policy-achieves-a-99-win-rate-against-katago/
…
MACHINE LEARNING
-
Adversarial Policy Defeats KataGo Go AI System 99%
By
–
-
Iterative approach to building and refining positive datasets
By
–
from my own experience you want something interactive and change your mind around quite a bit. so you're building the positive set, seeing the results, then tweaking your positive set over time. it's an incremental iterative thing.
-
Raw Scalable Intelligence: The Next AI Frontier
By
–
Pretty pictures are cool. But raw, scalable intelligence is on its way.
-
Beyond GPT-3: Next Wave of Superhuman AI Systems
By
–
In a couple of years, we’re going to look back at GPT-3 as a fun toy… The coming waves of AI systems will be smarter than humans in many ways. A different kind of intelligence. It’ll feel alien, yet familiar. And more generally capable than any single person alive today.
-
Hyperparameter Optimization Techniques and Practical Implementation Guide
By
–
Check out @subirmansukhani
’s technical blog on how to perform hyperparameter optimization. Then try the free accompanying Domino project yourself. https://
domino.buzz/3h6XxFq -
Artificial Intelligence Reshaping How We Shop Online
By
–
Artificial Intelligence Is Reshaping How We Shop https://
gearpatrol.com/style/a4184371
0/artificial-intelligence-shopping/
… @evanmalachosky @gearpatrol #AI #MachineLearning #DataScience #Robots #BigData #Analytics #100DaysofCode #IoT #serverless #womenwhocode #Python #DigitalTransformation #TensorFlow #DeepLearning #ecommerce -
GPU-Accelerated ML for Scalable Hybrid Cloud AI
By
–
Coming up: Need to use GPUs to accelerate ML? If you're at the AI Summit in Austin, check out Kjell Carlsson's talk TODAY, Nov. 2 from 3:05 – 3:30 CT, “Enabling a Scalable, Future-proof AI Ecosystem Across the Hybrid Cloud Enterprise." #MLOps #IoTWorld
-
GPT-3 Requires Chain-of-Thought for Problem Solving, Paper Analysis
By
–
Regardless how a human would do it, GPT-3 needs chain-of-thought to perform on these problems. Only skimmed the paper, but I don't see a specific deficit in ToM vs. known inability to do multi-hop inference without CoT.
-
Unlocking AI’s Full Potential for Human Intelligence
By
–
We're Not Using AI to Its Fullest Human Potential https://
time.com/6227118/eric-s
chmidt-ai-human-intelligence/
… @TIME #AI #MachineLearning #DataScience #BigData #Analytics #100DaysofCode #IoT #serverless #womenwhocode #DigitalTransformation #Robots #Python #TensorFlow #DeepLearning -
Adaptive Optimizers Should Track True Squared Gradients
By
–
Adaptive optimizers like Adam track the square of the gradient, but what they receive as the gradient is actually the sum of the gradients across the batch. It seems likely that better results at different hyper parameters could be obtained if backward passes emitted true grad^2.