@dvgnagy
 Tübingen, Germany
david-nagy.github.io
Joined February 2009
  • Tweets52
  • Following271
  • Followers127
David Nagy retweeted
Mark Ho@mark_ho_Jun 17
How do human minds make sense of big, messy problems? 😵‍💫🌀 How do we distill complexity into something simple enough to solve? 🤔💡 We’ll be tackling these questions (and more!) at two workshops on task representations, abstractions, and construals #CogSci2026 #CCN2026 🧵
1
5
2
48
2,622
Our @NatRevPsych Perspective with Gergo Orban and @TheCharleyWu is out. We take a normative perspective on the interplay of semantic and episodic memory, starting from the conundrum: how to learn a compression model while using it to compress the very data it will be updated on?
1
3
193
Full text accessible here: rdcu.be/epxF3
144
Looking forward to #CCN2023 tomorrow in Oxford! A great selection of research from @HanqiZhou_Ivy @alexandrawitt_ @AFurryBear1 @valerio_rubino @FraBraendle @RyutaroUchiyama @nir_moneta + many other hmc-lab.com members / collabs. See you there! @CogCompNeuro
7
1
39
6,994
We're wrapping up at #cshlNeuroAI 2022, and it's been wonderful to spend the week hearing about fascinating research in Neuro-AI. Such great interactions! Thanks to my co-organizers @TonyZador and @doristsao, and everyone else who participated. Hope to see you all next time!
2
5
3
58
From #PLOSCompBio: Optimal forgetting: Semantic compression of episodic memories Very relevant to the last #LearningSalon! dx.plos.org/10.1371/journal.…

Optimal forgetting: Semantic compression of episodic memories

Author summary Human memory performs surprisingly poorly in many everyday tasks, which have been richly documented in laboratory experiments. While constraints on memory resources necessarily imply a...

journals.plos.org
10
44
Check out & open review the @CogCompNeuro #CCNGAC proposal “What makes representations “useful”?” by @TBBake @wrongu @KriegeskorteLab @xaqlab, Alessandro Achille, Rosa Cao & Odelia Schwartz openreview.net/forum?id=HHk5…
1
10
17
David Nagy retweeted
EEML@EEMLcommunity24 Mar 2021
Deadline extended to *April 15* due to popular demand! List of speakers almost complete! Check out our website eeml.eu and submit your application ASAP!
2
15
2
80
David Nagy retweeted
EEML@EEMLcommunity1 Feb 2021
Exciting news! The (virtual) destination for EEML2021 is Budapest eeml.eu/! School dates July 7-15. Applications due March 31st! Don't miss out! Welcome David, @fhuszar, Gergo to the orgs team!
1
31
7
126
Our preprint on how lossy compression via generative models explains memory distortions in humans is now available on bioarxiv
bioRxiv Neuroscience@biorxiv_neursci8 May 2020
Optimal forgetting: Semantic compression of episodic memories biorxiv.org/cgi/content/shor… #biorxiv_neursci
1
11
Neuroscience professors. Can you help mentor neuromatch academy participants? We currently expect ~3k so we need help. Through matching you will work with students that are interesting to you. Please sign up here: docs.google.com/forms/d/e/1F…

The Neuromatch Academy - Mentor Application Round 1

What is The Neuromatch Academy? The Neuromatch Academy will be an open, online, 3-week intensive tutorial-based computational neuroscience training event (July 13-31, 2020). Participants from...

docs.google.com
2
39
7
56
#EEML2019 in Bucharest was a truly outstanding week. If you want to get to know more about my experience and some ML insights - checkout my new blog post! Massive thank you to @EEMLcommunity and all the wonderful people I got to know! roberttlange.github.io/posts…
4
5
28
Excellent new work by the Orban lab on stimulus-dependent noise correlations pnas.org/content/early/2019/…

Stimulus complexity shapes response correlations in primary visual cortex | PNAS

Spike count correlations (SCCs) are ubiquitous in sensory cortices, are characterized by rich structure, and arise from structured internal dynamic...

pnas.org
1
6
"BIVA: A Very Deep Hierarchy of Latent Variables for Generative Modeling". A skip-connected generative model and a bidirectional inference network are all you need! Work with @larsmaaloe @LieValentin @OleWinther1 arxiv.org/abs/1902.02102
2
26
91
Talks for the Probabilistic Programming conference #PROBPROG2018 are now available! youtube.com/playlist?list=PL… Includes, e.g., Zoubin Ghahramani @djsyclik Dave Blei @roydanroy @tom_rainforth @migorinova Stuart Russell, Josh Tenenbaum, and many others. probprog.cc/

PROBPROG 2018 Talks

Talks from the PROBPROG 2018 Conference, held October 5-6 2018 at the MIT Media Lab in Cambridge, MA.

youtube.com
1
77
4
270
Taming VAEs: A theoretical analysis of their properties and behaviour in the high-capacity regime. We also argue for a different way of training these models for robust control of key properties. It was fun thinking about this with @fabiointheuk arxiv.org/abs/1810.00597

Taming VAEs

In spite of remarkable progress in deep latent variable generative modeling, training still remains a challenge due to a combination of optimization and generalization issues. In practice, a...

arxiv.org
70
3
225
The Deep Learning and RL Summer School videos are up! videolectures.net/DLRLsummer…

Deep Learning and Reinforcement Learning Summer School, Toronto 2018

Deep neural networks are a powerful method for automatically learning distributed representations at multiple levels of abstraction. Over the past decade, they have dramatically pushed forward the...

videolectures.net
1
214
7
615
Our #cogsci2018 paper on compression of episodic memories is available at arxiv.org/abs/1806.07990, corresponding poster will be #154 on friday (27th)

Semantic Compression of Episodic Memories

Storing knowledge of an agent's environment in the form of a probabilistic generative model has been established as a crucial ingredient in a multitude of cognitive tasks. Perception has been...

arxiv.org
1
2
We organised a workshop at DALI 2018 on Goals and Principles of Representation Learning a few weeks ago. Videos here; including talks from @fhuszar @irinavlh @MatthiasBethge @jaschasd youtube.com/playlist?list=PL…

DALI 2018 - Workshop - Goals and Principles of Representation Learning

youtube.com
21
62
Introducing TensorFlow Probability: empowering ML researchers and practitioners to build sophisticated models quickly, leveraging state-of-the-art hardware Read about it on the TensorFlow blog ↓ goo.gl/1CQtU2

Introducing TensorFlow Probability

Posted by: Josh Dillon, Software Engineer; Mike Shwe, Product Manager; and Dustin Tran, Research Scientist — on behalf of the TensorFlow…

medium.com
2
321
27
713
Load more