HUANG Yifeng

Mathematician with a love of puzzling things. Postdoc at Emory. Former postdoc at USC, UBC. PhD '22 at U of Michigan.

[HUANG Yifeng @Y_F_Huang]

4/ We digested and internalized the proofs, isolated key ideas, and produced a human readable preprint. Structural problems inspired by the proof still remain; the problem is not "killed". Link to preprint: https://t.co/tkfRGdRcFS

[HUANG Yifeng @Y_F_Huang]

I am thrilled to share my recent preprint with Kenny Lau and @KenOno691 @axiommathai
We proved the HJO conjecture, a q-series identity arising from the geometry of singular curves, revealing new connections to cylindric partitions, compositional Dyck paths, and shuffle theorem.

[HUANG Yifeng @Y_F_Huang]

3/ Recent collaboration with Axiom AI has found new ingredients and proof strategy of the full HJO conjecture. Advanced techniques outside bijective counting are required: to prove HJO sum satisfies the same functional equation as the product side, we use the shuffle theorem.

[HUANG Yifeng @Y_F_Huang]

Great textbook. I knew of it when I taught at UBC, and I chose to use it again this semester.

[Math Cafe @Riazi_Cafe_en]

"Interactive Linear Algebra" (UBC Edition) https://t.co/RK26jpdmFL

[HUANG Yifeng @Y_F_Huang]

Not just an oraculous "this is true" verification, but also with canonicization in mind. Check out interactive https://t.co/jR4RbRaX9M, very interesting to play with.

[Axiom @axiommathai]

1/ We’re delighted to announce a milestone for AI-assisted mathematics:

With AxiomProver, we've completed a machine-checkable formalization of the “BGP246 theorem,” the best-known bound on recurring small gaps between primes.

Closest math has come to the Twin Prime Conjecture.

[HUANG Yifeng @Y_F_Huang]

Arxiv link: https://t.co/Cm1fKeUDsf. The full (3,b) case is now solved by https://t.co/JJLOPqHmTo

[HUANG Yifeng @Y_F_Huang]

We studied the first unknown family (3,b) of the HJO conjecture. We conjectured that it reduces to A_2 Rogers--Ramanujan identities of Warnaar. This assertion amounts to a sum=sum conjecture, which we prove for b at most 8. @axiommathai @leanprover @KenOno691

[HUANG Yifeng @Y_F_Huang]

We studied the first unknown family (3,b) of the HJO conjecture. We conjectured that it reduces to A_2 Rogers--Ramanujan identities of Warnaar. This assertion amounts to a sum=sum conjecture, which we prove for b at most 8. @axiommathai @leanprover @KenOno691

[HUANG Yifeng @Y_F_Huang]

Arxiv link: https://t.co/tBl0JCFB5K

[HUANG Yifeng @Y_F_Huang]

We computed the motive of high-rank Quot schemes of points over the X^a=Y^b singularities (with a,b coprime). Their generating function turns out to reveal brand new phenomena, including a conjectural bi-infinite family of Rogers--Ramanujan type identities (HJO conjecture).

[HUANG Yifeng @Y_F_Huang]

We computed the motive of high-rank Quot schemes of points over the X^a=Y^b singularities (with a,b coprime). Their generating function turns out to reveal brand new phenomena, including a conjectural bi-infinite family of Rogers--Ramanujan type identities (HJO conjecture).

[Andy Jiang @davikrehalt]

Trolling Fable be like ... (should I tell it that it's due to them?) {prompt written completely by GPT}

[HUANG Yifeng @Y_F_Huang]

Kudos to the A-team! The story originated from a machine-assisted research involving a series roughly like \sum_n q^{Q(n)} where n are integer vectors. But it won't make sense unless Q is positive definite. I gave a combinatorial proof, but the machine has to understand it.

[Carina Hong @CarinaLHong]

just finished something very cool ✨

we are also taking autoformalization requests from professional mathematicians - more to come soon 🪐

[HUANG Yifeng @Y_F_Huang]

Recent work with @axiommathai @KenOno691 involving a @leanprover autoformalization of a combinatorial proof of a positive definiteness theorem about a quadratic form. This is part of a bigger project in q-series. See https://t.co/irjwyYen4m

[HUANG Yifeng @Y_F_Huang]

Exciting theorem in the intersection of the geometry of singular curves and the combinatorics of numerical semigroups and Dyck paths! We prove that an explicit bilinear form on semigroup gaps gives a combinatorial statistic that extends dinv. The proof is autoformulated in Lean.

[HUANG Yifeng @Y_F_Huang]

Exciting theorem in the intersection of the geometry of singular curves and the combinatorics of numerical semigroups and Dyck paths! We prove that an explicit bilinear form on semigroup gaps gives a combinatorial statistic that extends dinv. The proof is autoformulated in Lean.

[HUANG Yifeng @Y_F_Huang]

@antfu7 拼音模式下有些韵会拆成介母+韵母（比如guan = g+u+an，如果答案里有hu这个字，输入guan的话看到u会匹配)，有些又不会（duo就还是d+uo, 不匹配u)，感觉太乱啦，规则里说只有声母和韵母没想到介母也要考虑