#RepresentationLearning
LightNews LightNews
globalcmd.bsky.social
Equation: Early+Mid+Late > Last-Layer Only 🧠
Fuse logits+states; rank tokens; verify steps; audit drift; calibrate outputs; cut hallucinations 🛠️ - GLCND.IO
Explore → https://glcnd.io/enhancing-llm-accuracy-by-leveraging-all-layers/
#AI #LLMs #RepresentationLearning
Enhancing LLM Accuracy by Leveraging All Layers - GLCND.IO
glcnd.io
September 18, 2025 at 9:16 AM
persagen.bsky.social
Geometric link: Convexity may bridge human & machine intelligence
phys.org/news/2025-07...

On convex decision regions in deep network representations
www.nature.com/articles/s41...

Convexity (algebraic geometry)
en.wikipedia.org/wiki/Convexi...

#ML #RepresentationLearning #convexity #DeepLearning
A geometric link: Convexity may bridge human & machine intelligence https://phys.org/news/2025-07-geometric-link-convexity-bridge-human.html On convex decision regions in deep network representations https://www.nature.com/articles/s41467-025-60809-y Convexity (algebraic geometry) https://en.wikipedia.org/wiki/Convexity_(algebraic_geometry) #ML #Representation Learning #convexity #AbstractAlgebra #DeepLearning
July 2, 2025 at 3:29 PM
neurreps.bsky.social
🧠✨ Join us for LAMINR: Learning and Aligning Manifolds of Single-Neuron Invariances — uncover the geometry of thought through implicit neural representations!

📅 Mon, Jun 16 @ 18:00 CET
🔗 Zoom | PW: 314159
🎙️ Part of the #NeurReps Global Seminar Series
#AI #Neuroscience #RepresentationLearning
June 13, 2025 at 7:48 AM
tum-aim-lab.bsky.social
#ISMRM 2025 took place this week, and here is a brief recap of our PhD students' presentations.

1/n
#CardiacMRI #ImplicitNeuralRepresentations #RepresentationLearning
May 18, 2025 at 5:22 PM
glenberseth.bsky.social
#reinforcementlearning #deeplearning #representationlearning
Lecture: youtu.be/-ZchWqayPR8
Robot Learning: Goal-Condition Panning
YouTube video by Montreal Robotics
youtu.be
March 25, 2025 at 5:01 PM
sqirllab.bsky.social
Recent work published by the #sqIRL Lab on the training of competitive deeper Forward-Forward Networks. #FF #localLearning #ML #RepresentationLearning
tmlr-pub.bsky.social TMLR Published Papers @tmlr-pub.bsky.social · Dec 13
The Trifecta: Three simple techniques for training deeper Forward-Forward networks

Thomas Dooms, Ing Jyh Tsang, Jose Oramas

Action editor: Pablo Castro

https://openreview.net/forum?id=a7KP5uo0Fp

#backpropagation #forward #deeper
December 15, 2024 at 4:56 PM
thilozimmermann.bsky.social
After #LunchBreak #FGML2021 will continue with
@laurent_besacie: #SelfSupervised #RepresentationLearning for Pre-training #SpeechSystems
#GérardBiau: Some Theoretical Insights into #Wasserstein #GANs
@KatharinaJutta Morik: Trustworthy #MachineLearning — The #CareLabel Approach
January 17, 2025 at 1:12 PM