Sign in

Paul Soulos

@paulsoulos.bsky.social
1.1K followers 383 following 21 posts

Computational Cognitive Science @JhuCogsci researching neurosymbolic methods. Previously wearable engineering @fitbit and @Google.

PostsRepliesMedia
Paul Soulos @paulsoulos.bsky.social · 30/05/2025
While both robotics and LM can be cast as next-token prediction, the token distribution for computer agents seems more like abstract motor programs (robotics) vs. language. This puts computer use on the trajectory of robotics which is slower than LLMs. 2/2
020
Paul Soulos @paulsoulos.bsky.social · 30/05/2025
Intriguing prediction from Trenton Bricken & @sholto-douglas.bsky.social on @dwarkesh.bsky.social's podcast: computer use agents "solved" in ~10 months 🖱️⌨️. This feels highly optimistic. I think that computer use is closer to robotics than language modeling. 1/2
120
Paul Soulos @paulsoulos.bsky.social · 11/12/2024
I’m presenting this work at 11a PT today in East Exhibit Hall at poster #4009. Come by and chat!
020
Paul Soulos @paulsoulos.bsky.social · 09/12/2024
📜 Check out our paper for all of the details and results. openreview.net/forum?id=fOQ....
openreview.net
Compositional Generalization Across Distributional Shifts with...
Neural networks continue to struggle with compositional generalization, and this issue is exacerbated by a lack of massive pre-training. One successful approach for developing neural systems which...
000
Paul Soulos @paulsoulos.bsky.social · 09/12/2024
📅 You can find me at the following presentations: - Poster Session 1 East #4009 on Wednesday, December 11, from 11a-2p PST. - System 2 Reasoning Workshop Spotlight Oral Talk on Sunday, December 15, from 9:30-10a PST. - System 2 Reasoning Workshop poster sessions on Sunday, December 15.
100
Paul Soulos @paulsoulos.bsky.social · 09/12/2024
📈DTM and sDTM operate on trees, and we introduce a very simple and dataset independent method to embed sequence inputs and outputs as trees. Across a variety of datasets and test time distributional shifts, sDTM outperforms fully neural and hybrid neurosymbolic models.
100
Paul Soulos @paulsoulos.bsky.social · 09/12/2024
🌳We introduce the Sparse Differentiable Tree Machine (sDTM), an extension of (DTM) that introduces a new way to represent trees in vector space. Sparse Coordinate Trees (SCT) reduce the parameter count and memory usage over the previous DTM by an order of magnitude and lead to a 30x speedup!
100
Paul Soulos @paulsoulos.bsky.social · 09/12/2024
Our previous work introducing the Differentiable Tree Machine (DTM) is an example of a unified neurosymbolic system where trees are represented and operated over in vector space.
100
Paul Soulos @paulsoulos.bsky.social · 09/12/2024
Hybrid systems use neural networks to parameterize symbolic components and can struggle with the same pitfalls as fully symbolic systems. In Unified Neurosymbolic systems, operations can simultaneously be viewed as either neural or symbolic, and this provides a fully neural path through the network.
100
Paul Soulos @paulsoulos.bsky.social · 09/12/2024
🧠 Neural networks struggle with compositionality, and symbolic methods struggle with flexibility and scalability. Neurosymbolic methods promise to combine the benefits of both methods, but there is a distinction between *hybrid* neurosymbolic methods and *unified* neurosymbolic methods.
100
Paul Soulos @paulsoulos.bsky.social · 09/12/2024
🚨 Thrilled to share that Compositional Generalization Across Distributional Shifts with Sparse Tree Operations received a spotlight award at #NeurIPS2024! 🌟 I'll present a poster on Tuesday and give an invited lightning talk at the System 2 Reasoning Workshop on Sunday. 🧵👇
1134
Paul Soulos @paulsoulos.bsky.social · 01/12/2024
Hi Melanie, I'll be there presenting some neurosymbolic work at the main conference and the system 2 workshop, as well as some other early work on Transformers and Computational Linguistics at the system 2 workshop! openreview.net/forum?id=fOQ... openreview.net/forum?id=6Pj...
openreview.net
Compositional Generalization Across Distributional Shifts with...
Neural networks continue to struggle with compositional generalization, and this issue is exacerbated by a lack of massive pre-training. One successful approach for developing neural systems which...
100
Paul Soulos @paulsoulos.bsky.social · 18/11/2024
Applied AGI scientist is a wild job title considering people have no idea how to even define AGI let alone what we should apply to create it.
060
Paul Soulos @paulsoulos.bsky.social · 13/11/2024
An important distinction that Sutskever makes in this article is that scale is not dead, but “Scaling the right thing matters more now than ever.” Vector symbolic architectures are a promising direction to scale symbolic methods in a fully differentiable manner. www.reuters.com/technology/a...
reuters.com
OpenAI and others seek new path to smarter AI as current methods hit limitations
Artificial intelligence companies like OpenAI are seeking to overcome unexpected delays and challenges in the pursuit of ever-biggerlarge language models by developing training techniques that use more human-like ways for algorithms to "think".
040
Reposted by Paul Soulos
xuan (ɕɥɛn / sh-yen) @xuanalogue.bsky.social · 11/11/2024
Okay the people requested one so here is an attempt at a Computational Cognitive Science starter pack -- with apologies to everyone I've missed! LMK if there's anyone I should add! go.bsky.app/KDTg6pv
7021991
Paul Soulos @paulsoulos.bsky.social · 11/11/2024
Want to learn more? Speak to the wonderful Kate McCurdy at #EMNLP2024 poster session 2, poster 1053, tomorrow from 11-12:30p ET. Check out the paper if you want the full details! aclanthology.org/2024.emnlp-m...
aclanthology.org
Toward Compositional Behavior in Neural Models: A Survey of Current Views
Kate McCurdy, Paul Soulos, Paul Smolensky, Roland Fernandez, Jianfeng Gao. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024.
010
Paul Soulos @paulsoulos.bsky.social · 11/11/2024
Researchers are split on HOW to achieve compositional behavior. Some propose data interventions, others argue we need entirely new model architectures, and some suggest we need to integrate symbolic paradigms.
100
Paul Soulos @paulsoulos.bsky.social · 11/11/2024
Key finding: ~75% of researchers agree that CURRENT neural models do NOT demonstrate true compositional behavior. Scale alone won't solve this - we need fundamental breakthroughs.
100
Paul Soulos @paulsoulos.bsky.social · 11/11/2024
We surveyed 79 top AI researchers about compositional behavior. Our goal? Map out the field's consensus and disagreements on how neural models process language to illuminate promising paths forward. Inspired by Dennett’s logical geography, we cluster participants by responses 🗺️
100
Paul Soulos @paulsoulos.bsky.social · 11/11/2024
Compositionality is fundamental to language: the ability to understand complex expressions by combining simpler parts. But do current AI models REALLY understand this? Spoiler: Most researchers say NO.
100
Paul Soulos @paulsoulos.bsky.social · 11/11/2024
I’m excited to share our survey investigating the current challenges and debates around achieving compositional behavior (CB) in language models, to be presented at #EMNLP2024! What makes language understanding truly intelligent? A thread unpacking our latest research 🤖📊🧵
110
Paul Soulos @paulsoulos.bsky.social · 19/09/2024
Besides being ergonomically beneficial, a split keyboard can prevent this from happening!
100