Mastodawn

Jialin Lu Dec 9, 2022

Pascal Hitzler Nov 30, 2022

We're starting a community slack for anybody interested in Neurosymbolic AI. (drivers include organizers of the annual workshop on the topic, and EiCs and EB members of the journal Neurosymbolic Artificial Intelligence that we're currently starting.
If you'd like to be on the slack, let me know (or anybody else who's already on it). You'll get an invite to your email then.

@tarekbesold @ArturGarcez @nesy_ai

Show thread

Jialin Lu Nov 22, 2022

However, the general circuit induction problem poses a more serious problem where no tricks I experimented provided any promising results. Essentially this is because it is optimizing a computation graph that is always changing in a non-smooth way, and the gradient just becomes to noisy to be useful.

Show thread

Jialin Lu Nov 22, 2022

A special parity chain task is discussed where relaxation approach is bound to fail when the problem size grows (as the number of local minima grows exponentially with the problem size).
Some tricks can do some help here (so we can solve this parity chain task reliably) （BTW the optimization process is also quite interesting visually)

Jialin Lu Nov 22, 2022

Updated some new thoughts regarding the TerpreT problem and my naive solution

https://luxxxlucy.github.io/projects/2021_terpret/index.html

The original TerpreT paper(https://arxiv.org/abs/1608.04428)
discussed solving program induction by gradient based optimization(after making the program differentiable by relaxation ).

#probprog #programsynthesis #neuralnetwork #deeplearning #NeuroSymbolic

Jialin Lu Nov 12, 2022

joaogui1 Nov 1, 2022

Papers accepted at #Neurips are now visible in OpenReview https://openreview.net/group?id=NeurIPS.cc/2022/Conference

NeurIPS 2022 Conference

Welcome to the OpenReview homepage for NeurIPS 2022 Conference

OpenReview

Jialin Lu Nov 9, 2022

Show thread

antonio vergari Nov 9, 2022

@emtiyaz @mathieualain

tl;dr: classical #PGMs can be compiled into #circuits #PCs and then #overparameterized to increase #expressiveness but retaining #tractability.

A gentle introduction for #probabilistic and #PGM folks is this paper https://web.cs.ucla.edu/~guyvdb/papers/ProbCirc20.pdf

A companion #video #tutorial is here https://www.youtube.com/watch?v=2RAG5-L9R70

I will do a new version of this tutorial with YooJung Choi and Robert Peharz at #NeurIPS2022 in a month!

Showing latest advancements in #reliable #inference with #PCs!

Jialin Lu Nov 8, 2022

M.L. Pacheco Nov 8, 2022

The slides and recording of our COLING 2022 tutorial on Neuro-Symbolic #NLProc are now available at: https://ns4nlp-coling.github.io

This is joint work with Dan Roth, Yejin Choi, Vivek Srikumar, Dan Goldwasser and Sean Welleck, all of whom I don't know how to find/tag on Mastodon :)

NS4NLP: Neuro-Symbolic Modeling for NLP

NS4NLP Coling 2022 Tutorial

NS4NLP: Neuro-Symbolic Modeling for NLP

Jialin Lu Nov 8, 2022

Jialin Lu Nov 7, 2022

Does anyone happen to know any mail-list/forum/discussion group for #neurosymbolic AI? I believe it is accumulating momentum gradually these days, yet I am having hard times finding relevant works.

Or perhap it is because the nature of neurosymbolic AI is just too heterogeneous, ill-defined and painfully diverse?

Jialin Lu Nov 8, 2022

antonio vergari Nov 6, 2022

I am a #Lecturer (Assistant Prof) in #ML at the University of #Edinburgh.

I research how #deep #probabilistic #models can be #reliable and can do #complex #reasoning with #guarantees in the real world and with #constraints.

Will toot about #automating #tractable #neuro #symbolic #AI and its #hype, #CS and #probabilistic #programming but also anything from #music to #politics and #fightsforrights.