Second Brain
Binxu Wang
Initializing search
Second Brain
Home
000 Zettelkasten
000 Zettelkasten
2D Convolutions
Ahead of Time (AOT) Compilation
Are less inductive biases better or worse?
Bit Palettization
Block Expansion
Convergence rate and Hessian spectra
Depthwise separable convolutions
Do Vision Foundation models exist?
Dynamic Resource Allocation in Vision
Effect of weight symmetries on training dynamics
Equivariance Initialization
Global Precedence Effect
Grokking
Group Axioms
Group direct product
Hardware specific structured pruning
Input dependent convolutions
K Means based Quantization
KV Cache
Linear Quantization
LoRa Adapter
Masked Image Modelling
Maximal pruning and functional recovery
Mean Attention Distance
Multiple global minima
Neural Network Quantization
Non translationally equivariant convolutions
Positive Logic Programs
Priors over Neural Network weights
PyTorch Functionalization
PyTorch Quantization for TensorRT
Rate Distortion and Spectral Analysis on Representations
Representation (Group Theory)
Residual stream
SK Centering
Should we separate semantics and pose?
100 Reference notes
100 Reference notes
101 Literature
101 Literature
4M Massively Multimodal Masked Modeling
A Brief Review of Hypernetworks in Deep Learning
A ConvNet for the 2020s
A Cookbook of Self Supervised Learning
A Hierarchy of Graph Neural Networks Based on Learnable Local Features
A Mathematical Framework for Transformer Circuits
A general theory of correct, incorrect, and extrinsic equivariance
A survey of quantization methods for efficient neural network inference
AWQ Activation aware Weight Quantization for LLM Compression and Acceleration
AdaGlimpse Active Visual Exploration with Arbitrary Glimpse Position and Scale
Adapting Vision Foundation Models for Plant Phenotyping
An Image is Worth More Than 16x16 Patches Exploring Transformers on Individual Pixels
An Investigation into Neural Net Optimization via Hessian Eigenvalue Density
An image is worth 16x16 words Transformers for image recognition at scale
Apple Intelligence Foundation Language Models
Approximately equivariant networks for imperfectly symmetric dynamics
Approximation Generalization Trade offs under (Approximate) Group Equivariance
Are Hierarchical Visual Representations Hierarchical
Autoequivariant Network Search via Group Decomposition
Average entropy of Gaussian mixtures
Battle of the Backbones A Large Scale Comparison of Pretrained Models across Computer Vision Tasks
Beyond cls Exploring the true potential of Masked Image Modeling representations
Block Transformer Global to Local Language Modeling for Fast Inference
BoxeR Box Attention for 2D and 3D Transformers
Building on Efficient Foundations Effectively Training LLMs with Structured Feedforward Layers
Byte Latent Transformer Patches Scale Better Than Tokens
CKConv Continuous Kernel Convolution For Sequential Data
Certifying Adapters Enabling and Enhancing the Certification of Classifier Adversarial Robustness
Color Equivariant Convolutional Networks
Color Space Transformation Network
ConViT Improving Vision Transformers with Soft Convolutional Inductive Biases
Correlational Image Modeling for Self Supervised Visual Pre Training
Curiosity driven Exploration by Self supervised Prediction
DETRs Beat YOLOs on Real time Object Detection
DETRs with Collaborative Hybrid Assignments Training
DINOv2 Learning Robust Visual Features without Supervision
Deep Learning Book
Deep Learning is Not So Mysterious or Different
Deformable Convolutional Networks
Deformable DETR Deformable Transformers for End to End Object Detection
DeiT III Revenge of the ViT
Dense Contrastive Learning for Self Supervised Visual Pre Training
DenseNets Reloaded Paradigm Shift Beyond ResNets and ViTs
Discovering Symmetry Breaking in Physical Systems with Relaxed Group Convolution
DropPos Pre Training Vision Transformers by Reconstructing Dropped Positions
EA ViT Efficient Adaptation for Elastic Vision Transformer
EVA 02 A Visual Representation for Neon Genesis
Early Convolutions Help Transformers See Better
Efficient Equivariant Transfer Learning from Pretrained Models
Efficient Modulation for Vision Networks
EfficientViT SAM Accelerated Segment Anything Model Without Accuracy Loss
Emergent Equivariance in Deep Ensembles
Emerging Properties in Self Supervised Vision Transformers
End to End Object Detection with Transformers
Equi Tuning Group Equivariant Fine Tuning of Pretrained Models
Equivariance with Learned Canonicalization Functions
Equivariance aware architectural optimization of neural networks
Equivariant Representation Learning via Class Pose Decomposition
Exact Conversion of In Context Learning to Model Weights in Linearized Attention Transformers
Exploiting Redundancy Separable Group Convolutional Networks on Lie Groups
Exploring Plain Vision Transformer Backbones for Object Detection
Exploring Simple Siamese Representation Learning
FLSL Feature level Self supervised Learning
FNet Mixing Tokens with Fourier Transforms
Fast, Expressive SE(n) Equivariant Networks through Weight Sharing in Position Orientation Space
Fixing the train test resolution discrepancy
FlexTok Resampling Images into 1D Token Sequences of Flexible Length
FlexiViT One Model for All Patch Sizes
Fourier Transformer Fast Long Range Modeling by Removing Sequence Redundancy with FFT Operator
Franca Nested Matryoshka Clustering for Scalable Visual Representation Learning
From Pixels to Components Eigenvector Masking for Visual Representation Learning
G SGD Optimizing ReLU Neural Networks in its Positively Scale Invariant Space
Grokked Transformers are Implicit Reasoners A Mechanistic Journey to the Edge of Generalization
Guillotine Regularization Why removing layers is needed to improve generalization in Self Supervised Learning
Harmonics of Learning Universal Fourier Features Emerge in Invariant Networks
HoPE A Novel Positional Encoding Without Long Term Decay for Enhanced Context Awareness and Extrapolation
How Does SimSiam Avoid Collapse Without Negative Samples A Unified Understanding with Self supervised Contrastive Learning
How JEPA Avoids Noisy Features The Implicit Bias of DeepLinear Self Distillation Networks
How do vision transformers work?
How to represent part whole hierarchies in a neural network
Hydra Bidirectional State Space Models Through Generalized Matrix Mixers
HydraViT Stacking Heads for a Scalable ViT
Hyperspherical Variational Auto Encoders
Improving Convergence and Generalization Using Parameter Symmetries
Improving Self Consistency in LLMs through Probabilistic Tokenization
In Search of Projectively Equivariant Networks
Into the Rabbit Hull From Task Relevant Concepts in DINO to Minkowski Geometry
Intuitive physics understanding emerges from self supervised pretraining on natural videos
Is ImageNet worth 1 video Learning strong image encoders from 1 long unlabelled video
KV Cache Steering for Inducing Reasoning in Small Language Models
Knowledge Transfer from Vision Foundation Models for Efficient Training of Small Task specific Models
LRP QViT Mixed Precision Vision Transformer Quantization via Layer wise Relevance Propagation
Layer by Layer Uncovering Hidden Representations in Language Models
Learned Gridification for Efficient Point Cloud Processing
Learning Continually by Spectral Regularization
Learning Partial Equivariances from Data
Learning Representations on the Unit Sphere Investigating Angular Gaussian and von Mises Fisher Distributions for Online Continual Learning
Learning Unseen Modality Interaction
Learning both Weights and Connections for Efficient Neural Networks
Learning to Learn without Forgetting using Attention
Learning with Unmasked Tokens Drives Stronger Vision Learners
LieRE Generalizing Rotary Position Encodings
Llama 2 Open Foundation and Fine Tuned Chat Models
LoRA Low Rank Adaptation of Large Language Models
LoRA vs Full Fine tuning An Illusion of Equivalence
Location Aware Self Supervised Transformers for Semantic Segmentation
Lost in Time A New Temporal Benchmark for VideoLLMs
Low Resource Vision Challenges for Foundation Models
Mamba Linear Time Sequence Modeling with Selective State Spaces
Masked Autoencoders Are Scalable Vision Learners
MatFormer Nested Transformer for Elastic Inference
Matryoshka Representation Learning
Memorization Through the Lens of Curvature of Loss Function Around Samples
Mixture of LoRa Experts
MoSiC Optimal Transport Motion Trajectory for Dense Self Supervised Learning
MobileCLIP Fast Image Text Models through Multi Modal Reinforced Training
MobileViT light weight, general purpose, and mobile friendly vision transformer
Model Compression in Practice Lessons Learned from Practitioners Creating On device Machine Learning Experiences
Near, far Patch ordering enhances vision foundation models' scene understanding
NeoBabel A Multilingual Open Tower for Visual Generation
Neural Mechanics Symmetry and Broken Conservation Laws in Deep Learning Dynamics
No Train, all Gain Self Supervised Gradients Improve Deep Frozen Representations Yuki M Asano
Occam's model Selecting simpler representations for better transferability estimation
On Good Practices for Task Specific Distillation of Large Pretrained Visual Models
On the Relationship between Self Attention and Convolutional Layers
On the Symmetries of Deep Learning Models and their Internal Representations
On the duality between contrastive and non contrastive self supervised learning
OpenELM An Efficient Language Model Family with Open source Training and Inference Framework
Optimal Brain Damage
Optimization Dynamics of Equivariant and Augmented Neural Networks
PART Self supervised Pretraining with Pairwise Relative Translations
PIN Positional Insert Unlocks Object Localisation Abilities in VLMs
Parameter Efficient Fine tuning of Self supervised ViTs without Catastrophic Forgetting
Parameter Efficient Fine Tuning for Pre Trained Vision Models A Survey
Patch Wise Self Supervised Visual Representation Learning A Fine Grained Approach
PatchRot A Self Supervised Technique for Training Vision Transformers
Position Prediction as an Effective Pretraining Strategy
Principal Components Enable A New Language of Images
Progress measures for grokking via mechanistic interpretability
Provably Strict Generalisation Benefit for Equivariant Models
ProxylessNAS Direct Neural Architecture Search on Target Task and Hardware
R MAE Regions Meet Masked Autoencoders
Rate–Distortion–Perception Trade Off in Information Theory, Generative Models, and Intelligent Communications
Refusal in Language Models Is Mediated by a Single Direction
Relaxed Octahedral Group Convolution for Learning Symmetry Breaking in 3D Physical Systems
Relaxing Equivariance Constraints with Non stationary Continuous Filters
Rethinking Lossy Compression The Rate Distortion Perception Tradeoff
Retrospective EIE Efficient Inference Engine onSparse and Compressed Neural Network
Revealing the Utilized Rank of Subspaces of Learning in Neural Networks
Revisiting Feature Prediction for Learning Visual Representations from Video
Rewrite the Stars
Robust Self Supervised Learning with Lie Groups
Rotary Position Embedding for Vision Transformer
Round and Round We Go! What makes Rotary Positional Encodings useful?
SAM CLIP Merging Vision Foundation Models towards Semantic and Spatial Understanding
Scaling (Down) CLIP A Comprehensive Analysis of Data, Architecture, and Training Strategies
Scaling and Benchmarking Self Supervised Visual Representation Learning
Searching for Efficient Linear Layers over a Continuous Space of Structured Matrices
Segment Anything
Self Guided Diffusion Models
Self Supervised Detection of Perfect and Partial Input Dependent Symmetries
Self Supervised Learning from Images with a Joint Embedding Predictive Architecture
Self Supervised Learning of Object Parts for Semantic Segmentation
Self supervised learning of Split Invariant Equivariant representations
Self supervised learning of intertwined content and positional features for object detection
SimPLR A Simple and Plain Transformer for Scaling Efficient Object Detection and Segmentation
Simplifying DINO via Coding Rate Regularization
Simultaneous linear connectivity of neural networks modulo permutation
Slicing Vision Transformer for Flexible Inference
Stand Alone Self Attention in Vision Models
Stochastic positional embeddings improve masked image modeling
Structuring Representation Geometry with Rotationally Equivariant Contrastive Learning
Surgical Fine Tuning Improves Adaptation to Distribution Shifts
Surgical DINO Adapter Learning of Foundation Models for Depth Estimation in Endoscopic Surgery
Symmetries in Overparametrized Neural Networks A Mean Field View
Talaria Interactively Optimizing Machine Learning Models for Efficient Inference
The Empirical Impact of Neural Parameter Symmetries, or Lack Thereof
The Forward Forward Algorithm Some Preliminary Investigations
The Hidden Uniform Cluster Prior in Self Supervised Learning
The Lie derivative for measuring learned equivariance
The Unreasonable Ineffectiveness of the Deeper Layers
Three things everyone should know about Vision Transformers
TiC CLIP Continual Training of CLIP models
To Compress or Not to Compress Self Supervised Learning and Information Theory A Review
Toward a Geometrical Understanding of Self supervised Contrastive Learning
Towards Understanding the Spectral Bias of Deep Learning
Towards a Definition of Disentangled Representations
Training quantized nets A deeper understanding
Understanding Deep Learning Chapter 10
Understanding Deep Learning Chapter 20
Understanding symmetries in deep networks
Unlocking Slot Attention by Changing Optimal Transport Costs
Unsupervised Learning of Visual Features by Contrasting Cluster Assignments
Unsupervised Visual Representation Learning by Context Prediction
Using Degeneracy in the Loss Landscape for Mechanistic Interpretability
VICRegL Self Supervised Learning of Local Visual Features
Variance Covariance Regularization Enforces Pairwise Independence in Self Supervised Representations
ViDT An Efficient and Effective Fully Transformer based Object Detector
ViTAR Vision Transformer with Any Resolution
Vision Mamba Efficient Visual Representation Learning with Bidirectional State Space Model
Vision Transformer with Deformable Attention
Vision Transformers Need Registers
What Do Self Supervised Vision Transformers Learn?
nGPT Normalized Transformer with Representation Learning on the Hypersphere
102 Authors
102 Authors
Adrien Bardes
Albert Gu
Alex Flinth
Alex Vitvitskyi
Alexander Kirillov
Alexey Dosovitskiy
Amir Zamir
Ananya Kumar
Andreas Loukas
Andreas Savakis
Andrew Gordon Wilson
Angela Fan
Annie S. Chen
Antonio Orvieto
Ardavan Pedram
Armand Joulin
Attila Lengyel
Binxu Wang
Boris Ginsburg
Boshi Wang
Byeongho Heo
Caglar Gulcehre
Carmen Amo Alonso
Cees G. M. Snoek
Chelsea Finn
Cheng Ping Hsieh
Chong Wang
Christopher Olah
Christos Perivolaropoulos
Daniel M. Roy
Daniel Ulbricht
David M. Knigge
David W. Romero
David W. Zhang
Dheeraj Varghese
Diane Larlus
Donghyun Kim
Dongyoon Han
Duy Kien Nguyen
Edward J. Hu
Edward Z. Yang
Efstratios Gavves
Eric Mintun
Erik J. Bekkers
Eshan Verma
Fahim Tajwar
Fartash Faghri
Federico Barbero
Florian Bordes
Francisco Massa
Fred Hohman
Furu Wei
Gabriel Synnaeve
Geoffrey Hinton
Gertjan Burghouts
Gertjan J. Burghouts
Gintare Karolina Dziugaite
Giovanni Chierchia
Grégoire Mialon
Hadi Pouransari
Han Cai
Hanzi Mao
Haoxiang Wang
Hervé Jegou
Huaxiu Yao
Hugo Touvron
Huizi Mao
Ilya Loshchilov
Ioana Simion
Isha Garg
Ishan Misra
Jan E. Gerken
Javier Maass Martinez
Jean Ponce
Jean Baptiste Cordonnier
Jean François Bercher
Jeff Pool
Jesse Cai
Jing Pu
Joaquin Fontbona
Joaquin Vanschoren
John Denker
John Tran
Julien Mairal
Juliette Marrie
Kaiming He
Kamyar Azizzadenesheli
Kaushik Roy
Laurent Najman
Lawrence Chan
Lucas Beyer
Lucia Donatelli
Lucius Bushnaq
Maciej Wołczyk
Mahdi Soltanolkotabi
Mahmoud Assran
Marc Finzi
Mark A. Horowitz
Mark Ibrahim
Martin Jaggi
Martin R. Oswald
Marzieh Fadaee
Mathilde Caron
Max Belistky
Maxime Oquab
Mehrdad Farajtabar
Micah Goldblum
Michael Arbel
Michael Dorkenwald
Mohammad Mahdi Derakhshani
Mohammad Rastegari
Mohammadreza Salehi
Namuk Park
Navin Ranjan
Neel Nanda
Neil Houlsby
Nicolas Carion
Nicolas Michel
Nicolas Usunier
Olivier J. Hénaff
Oncel Tuzel
Pascal Vincent
Patrick Forré
Pavan Kumar Anasosalu Vasu
Percy Liang
Petar Veličković
Phillip Isola
Piotr Bojanowski
Piotr Dollár
Quentin Garrido
Randall Balestriero
Ravid Shwartz Ziv
Raviteja Vemulapalli
Razvan Pascanu
Robin Walters
Romain Cosentino
Romain Negrel
Rose Yu
Ross Girshick
Rui Wang
Ruoming Pang
Sachin Mehta
Saining Xie
Sangdoo Yun
Sanghyuk Chun
Sara Solla
Sergey Zagoruyko
Shaohan Huang
Shashanka Venkataramanan
Sibylle Hess
Simeng Sun
Simon J.D. Prince
Skander Moalla
Soham De
Song Han
Song Park
Songkuk Kim
Sonia Joseph
Sourya Basu
Stéphane d'Ascoli
Sukjun Hwang
Taekyung Kim
Tete Xiao
Thomas Fel
Thomas Kipf
Tim R. Davidson
Tom Goldstein
Tom Gunter
Tom Lieberum
Tongzhou Wang
Vaibhav Aggarwal
Valentinos Pariza
Walter Simoncini
Wei Lu
William J. Dally
Wonjae Kim
Xiang Yue
Xingyu Liu
Xinlei Chen
Xiuying Wei
Xu Ma
Xun Wu
Yanghao Li
Yann LeCun
Yelong Shen
Yi Ma
Yoonho Lee
Yubei Chen
Yuki M. Asano
Zeyuan Allen Zhu
Zhuoyang Zhang
Ziaoyi Zhang
Zirui Wang
Ziyang Wu
103 Affiliations
103 Affiliations
Anthropic
Apollo Research
Apple
CLAIRE
Carnegie Mellon University
Chalmers University of Technology
Cohere
EPFL
FAIR
FunAI Lab
Google DeepMind
Google
HKU
IBM Research
INRIA
MIT
McGill University
Meta
Microsoft
Mila Quebec AI Institute
NVIDIA
Naver AI Lab
Naver Cloud AI
Naver Labs Europe
New York University
Northeastern University
OpenAI
Princeton University
PyTorch
QUVA Lab
Rochester Institute of Technology
Stanford
TNO
TU Delft
Tsinghua University
UC Berkeley
UC San Diego
UC Santa Barbara
UCLA
Univ Gustave Eiffel
University of Amsterdam
University of Chile
University of Illinois at Urbana Champaign
University of Oxford
University of Technology Nuremberg
VIS Lab
Vector Institute
Vrije Universiteit Amsterdam
Yonsei University
104 Other
104 Other
EPFL CS439 Optimization for Machine Learning
GPU mode Sparsity
Introducing Apple’s On Device and Server Foundation Models
Introduction to Quantization on PyTorch
Let's talk about the Python Dispatcher
MIT 65940 TinyML and Efficient Deep Learning Computing
Optimizing Vision Transformer Model for Deployment
PyTorch ExecuTorch Export IR Specification
PyTorch ExecuTorch How ExecuTorch works?
PyTorch ExecuTorch Quantization Overview
PyTorch Functionalization in PyTorch Everything you need to know
PyTorch PyTorch 2 Export Post Training Quantization
PyTorch Quantization
PyTorch Compilers What makes PyTorch beloved makes it hard to compile
PyTorch Conference 2024 Fast Sparse Vision Transformers with minimal accuracy loss
PyTorch Conference 2024 What’s new in torch.export?
PyTorch Conference 2024
PyTorch Eager Mode Quantization TensorRT Acceleration
PyTorch internals
Quantized Transfer Learning for Computer Vision Tutorial
Reinforcement Learning An Introduction Chapter 10
Reinforcement Learning An Introduction Chapter 11
Reinforcement Learning An Introduction Chapter 13
Reinforcement Learning An Introduction Chapter 16
Reinforcement Learning An Introduction Chapter 2
Reinforcement Learning An Introduction Chapter 3
Reinforcement Learning An Introduction Chapter 4
Reinforcement Learning An Introduction Chapter 5
Reinforcement Learning An Introduction Chapter 6
Reinforcement Learning An Introduction Chapter 7
Reinforcement Learning An Introduction Chapter 8
Reinforcement Learning An Introduction Chapter 9
Reinforcement Learning An Introduction
TinyML and Efficient Deep Learning Computing Lecture 12
TinyML and Efficient Deep Learning Computing Lecture 3
TinyML and Efficient Deep Learning Computing Lecture 5
TinyML and Efficient Deep Learning Computing Lecture 6
TinyML and Efficient Deep Learning Computing
Tweet Stable Diffusion XL on iPhone with Core ML!
Binxu Wang
Properties
affiliation
Harvard, Kempner Institute