“I found this to be an interesting read, though it's just a high-level plan, vs. breaking new ground. I like how the ideas are synthesized. A fun detail is to be found in a footnote: "A mathematician is a person who can find analogies between theorems; a better mathematician is one who can see analogies between proofs and the best mathematician can notice…”
Machine learning and information theory concepts towards an AI Mathematician
The current state-of-the-art in artificial intelligence is impressive, especially in terms of mastery of language, but not so much in terms of mathematical reasoning. What could be missing? Can we learn something useful about that gap from how the brains of mathematicians go about their craft? This essay builds on the idea that current deep learning mostly succeeds at system 1 abilities -- which correspond to our intuition and habitual behaviors -- but still lacks something important regarding system 2 abilities -- which include reasoning and robust uncertainty estimation. It takes an information-theoretical posture to ask questions about what constitutes an interesting mathematical statement, which could guide future work in crafting an AI mathematician. The focus is not on proving a given theorem but on discovering new and interesting conjectures. The central hypothesis is that a desirable body of theorems better summarizes the set of all provable statements, for example by having a small description length while at the same time being close (in terms of number of derivation steps) to many provable statements.
候选层:至少 2 个不同的人、在 2 个不同月份提到过,还没到主书架的 3 人 3 个月。没有做立场标注,所以只写“被提及”。
HN 书评
没有评论直接链到它;下面是它被提交的讨论串里回复最多的评论(不计入统计)“A key aspect in AlphaProof was how to find good training problems. For AlphaProof they managed to scrape 1 million "International Mathematical Olympiad (IMO) style" problems. Here Bengio seems to just suggest using estimated entropy `E_P(t|T(S))[− log P(t | T(S))]` of the theorem to find interesting/surprising theorems to try and prove. This reminds me …”
“I love that "system 1" and "system 2" are tossed out as fundamental concepts of mind when they're basically labels for some semi-random ideas based on bullshit studies in a pop sci fluff book. Is there any actual science that ties these ideas to mathematical reasoning? Then we have: >> By analogy, we can now state the main hypothesis proposed in this pa…”
- Machine learning and information theory concepts towards an AI Mathematician ↗carlossouza · 2024-03-08 · 3 分
- Machine learning and information theory concepts towards an AI Mathematician ↗marojejian · 2024-10-12 · 109 分