Large Language Models Can Self-Improve Huang et al.: https://
arxiv.org/abs/2210.11610 #ArtificialIntelligence #DeepLearning #MachineLearning
RESEARCH
-
Large Language Models Demonstrate Self-Improvement Capabilities
By
–
-
Brain-Immune System Axis: Three Rheostats for Control
By
–
Featuring 3 different rheostats that can dial up or down the brain-body immune system axis. One of them is from this week's @ScienceMagazine https://
science.org/doi/10.1126/sc
ience.adl4237
… https://
erictopol.substack.com/p/guardians-of
-the-brain
… -
Search and Deep Learning: The Future of AI Systems
By
–
Search is a fully general technique. But program search today is at the level of sophistication that deep learning was at in 2011. Hence why deep learning appears more general. Soon, you will see combined search + deep learning systems vastly outperform anything you can do with
-
Superintelligence Gates Breaking Through Reality Barrier
By
–
The Gates Of Superintelligence Are BREACHING Reality. $ASI #ASIFirst
-
AI researcher persecution drives innovation to other countries
By
–
They'd move to Singapore. Killing AI researchers in your own country does not make your own country safe. It means some other country spawns the horror that kills humanity instead.
-
Verifying AI Model Outputs and Public Signatures Without Revealing Weights
By
–
The use case is if you have a lot of people who all want to verify that the same known large expensive AI model said a thing. Another use case might be if there's some way to have a fixed model public signature without revealing its weights.
-
ARC-AGI: Non-Deep Learning Achieves 40% on Private Test Set
By
–
On what, ARC-AGI? 40% with no deep learning at all right now (private test set). Best DL based solution is at 55%.
-
Comparing iterative reasoning systems versus discrete program search efficiency
By
–
Re: "optimal way to do reasoning". The idea here is that if you have an "iterated system 1" type reasoning system that takes ~$1k and O(hour) in compute to find a correct program to solve a simple task, vs. a discrete program search system that does the same in 0.1s on your
-
WordPress categories focused on AI topics
By
–
For anyone else reading along, tested this idea in a new post:
-
Discussion of CoT and model memorization/novelty
By
–
Also, the reason for the light square condition is to inject arbitrary novelty, adding credence the solution isn’t memorized from training I’m not worried it is; in that 5m45s o1 spews out a dozen pages of CoT summary broadly consistent with its answer — representative excerpt: