[DL Math+Efficiency] Yedi Zhang - Saddle-to-Saddle Dynamics Explains A Simplicity Bias Across NNs
Embedded AI Lab @TUG
[DL Math+Efficiency] Yedi Zhang - Saddle-to-Saddle Dynamics Explains A Simplicity Bias Across NNs
52:14
[DL Math+Efficiency] Rahim Entezari - Fast Video Generation
Embedded AI Lab @TUG
[DL Math+Efficiency] Rahim Entezari - Fast Video Generation
47:52
[DL Math+Efficiency] Ting Han - Flatness is Necessary, Neural Collapse is Not
Embedded AI Lab @TUG
[DL Math+Efficiency] Ting Han - Flatness is Necessary, Neural Collapse is Not
25:26
[DL Math+Efficiency] Flavio Martinelli - Flat Channels to Infinity in Neural Loss Landscapes
Embedded AI Lab @TUG
[DL Math+Efficiency] Flavio Martinelli - Flat Channels to Infinity in Neural Loss Landscapes
37:04
[DL Math+Efficiency] Georgios Vlassis - Beyond Outliers: A Study of Optimizers Under Quantization
Embedded AI Lab @TUG
[DL Math+Efficiency] Georgios Vlassis - Beyond Outliers: A Study of Optimizers Under Quantization
33:44
[DL Math+Efficiency] Tom Jacobs - Mask in the Mirror: Implicit Sparsification
Embedded AI Lab @TUG
[DL Math+Efficiency] Tom Jacobs - Mask in the Mirror: Implicit Sparsification
18:11
[DL Math+Efficiency] Giulia Lanzillotta - Testing knowledge distillation theories with dataset size
Embedded AI Lab @TUG
[DL Math+Efficiency] Giulia Lanzillotta - Testing knowledge distillation theories with dataset size
57:37
[DL Math+Efficiency] Shuaicheng Niu - Towards Versatile Test-Time Adaptation
Embedded AI Lab @TUG
[DL Math+Efficiency] Shuaicheng Niu - Towards Versatile Test-Time Adaptation
34:30
[DL Math+Efficiency] Oscar Skean - Uncovering Hidden Representations in Language Models
Embedded AI Lab @TUG
[DL Math+Efficiency] Oscar Skean - Uncovering Hidden Representations in Language Models
35:06
[EfficientML] Abhishek Tyagi - Dynamic Sparse Training of Diagonally Sparse Networks
Embedded AI Lab @TUG
[EfficientML] Abhishek Tyagi - Dynamic Sparse Training of Diagonally Sparse Networks
23:36
[EfficientML] Simla Burcu Harma - Effective Interplay between Sparsity and Quantization
Embedded AI Lab @TUG
[EfficientML] Simla Burcu Harma - Effective Interplay between Sparsity and Quantization
23:06
[EfficientML] Dong Wang: Forget the Data and Fine-Tuning! Just Fold the Network to Compress
Embedded AI Lab @TUG
[EfficientML] Dong Wang: Forget the Data and Fine-Tuning! Just Fold the Network to Compress
23:09
Driver Drowsiness Detection App by Elin Ruud and Daniel Kreimer
Embedded AI Lab @TUG
Driver Drowsiness Detection App by Elin Ruud and Daniel Kreimer
1:19
Circuit Detector Demo by Martin Schachl and Michael Zweimüller
Embedded AI Lab @TUG
Circuit Detector Demo by Martin Schachl and Michael Zweimüller
1:24
Jump and Run Game by Fabian Treu and David Prutej
Embedded AI Lab @TUG
Jump and Run Game by Fabian Treu and David Prutej
1:54
[EfficientML] Benjamin Vandersmissen - Improving NN Accuracy by Training with a Twin Network
Embedded AI Lab @TUG
[EfficientML] Benjamin Vandersmissen - Improving NN Accuracy by Training with a Twin Network
31:25
[EfficientML] Bingcong Li - On the Crucial Role of Initialization for Matrix Factorization
Embedded AI Lab @TUG
[EfficientML] Bingcong Li - On the Crucial Role of Initialization for Matrix Factorization
37:27
[EfficientML] Lorenzo Sani - Photon: Federated LLM Pre-Training
Embedded AI Lab @TUG
[EfficientML] Lorenzo Sani - Photon: Federated LLM Pre-Training
41:19
[EfficientML] Ankit Sonthalia: Do Deep Neural Network Solutions Form a Star Domain?
Embedded AI Lab @TUG
[EfficientML] Ankit Sonthalia: Do Deep Neural Network Solutions Form a Star Domain?
20:04
[EfficientML] Ramón Calvo González - Leveraging the True Depth of LLMs
Embedded AI Lab @TUG
[EfficientML] Ramón Calvo González - Leveraging the True Depth of LLMs
18:23
[EfficientML] Denis Korzhenkov - Mobile Video Diffusion
Embedded AI Lab @TUG
[EfficientML] Denis Korzhenkov - Mobile Video Diffusion
27:05
[EfficientML] Shiwei Liu - The Curse of Depth in Large Language Models
Embedded AI Lab @TUG
[EfficientML] Shiwei Liu - The Curse of Depth in Large Language Models
40:19
[EfficientML] Janek Haberer - HydraViT: Stacking Heads for a Scalable ViT
Embedded AI Lab @TUG
[EfficientML] Janek Haberer - HydraViT: Stacking Heads for a Scalable ViT
31:30
[EfficientML] Alexander Theus and Olin Geimer: Towards Meta-Pruning via Optimal Transport
Embedded AI Lab @TUG
[EfficientML] Alexander Theus and Olin Geimer: Towards Meta-Pruning via Optimal Transport
26:35
[EfficientML] Nils Walter: The Uncanny Valley: Adversarial Robustness from a Flatness Perspective
Embedded AI Lab @TUG
[EfficientML] Nils Walter: The Uncanny Valley: Adversarial Robustness from a Flatness Perspective
27:02
[EfficientML] Eldar Kurtic: Mathador-LM: A Dynamic Benchmark for Mathematical Reasoning on LLMs
Embedded AI Lab @TUG
[EfficientML] Eldar Kurtic: Mathador-LM: A Dynamic Benchmark for Mathematical Reasoning on LLMs
25:39
[EfficientML] Francesco Tosoni: Toward Greener Matrix Operations by Lossless Compressed Formats
Embedded AI Lab @TUG
[EfficientML] Francesco Tosoni: Toward Greener Matrix Operations by Lossless Compressed Formats
40:42
[EfficientML] Ke Wang & Nikolaos Dimitriadis: Localizing Task Information for Improved Model Merging
Embedded AI Lab @TUG
[EfficientML] Ke Wang & Nikolaos Dimitriadis: Localizing Task Information for Improved Model Merging
30:54
[EfficientML] Flavio Martinelli - Expand-and-Cluster: Parameter Recovery of Neural Networks
Embedded AI Lab @TUG
[EfficientML] Flavio Martinelli - Expand-and-Cluster: Parameter Recovery of Neural Networks
46:16
[EfficientML] Dong Wang - Subspace-Configurable Networks
Embedded AI Lab @TUG
[EfficientML] Dong Wang - Subspace-Configurable Networks
22:27
[EfficientML] Peng Sun - Efficiency for Free: Ideal Data Are Transportable Representations
Embedded AI Lab @TUG
[EfficientML] Peng Sun - Efficiency for Free: Ideal Data Are Transportable Representations
33:40
[MCL] Master Wang Game: Human Activity Recognition and Interaction
Embedded AI Lab @TUG
[MCL] Master Wang Game: Human Activity Recognition and Interaction
0:36
[MCL] TopOut App: AI-Powered Climbing Recording
Embedded AI Lab @TUG
[MCL] TopOut App: AI-Powered Climbing Recording
0:48
[EfficientML] Da Xiao - Improving Transformers with Dynamically Composable Multi-Head Attention
Embedded AI Lab @TUG
[EfficientML] Da Xiao - Improving Transformers with Dynamically Composable Multi-Head Attention
40:46
[EfficientML] Keller Jordan -- 94% on CIFAR-10 in 3.29 Seconds on a Single GPU
Embedded AI Lab @TUG
[EfficientML] Keller Jordan -- 94% on CIFAR-10 in 3.29 Seconds on a Single GPU
31:37
[EfficientML] Diganta Misra -- Just Say the Name: Online Continual Learning via Data Generation
Embedded AI Lab @TUG
[EfficientML] Diganta Misra -- Just Say the Name: Online Continual Learning via Data Generation
54:21
[MCL] Wireless DrumKit App by Markus-Philipp Gherman and Fabian Moik
Embedded AI Lab @TUG
[MCL] Wireless DrumKit App by Markus-Philipp Gherman and Fabian Moik
0:55
[MCL] Indoor Localization by Oliver Kreuzer and Thomas Stranger
Embedded AI Lab @TUG
[MCL] Indoor Localization by Oliver Kreuzer and Thomas Stranger
1:16
[Embedded ML] Deep Model Inference Demo on STM32 Discovery Kit by Peter Prodinger
Embedded AI Lab @TUG
[Embedded ML] Deep Model Inference Demo on STM32 Discovery Kit by Peter Prodinger
3:09
[Embedded ML] Magic Wand Demo 2021 by Rahim Entezari
Embedded AI Lab @TUG
[Embedded ML] Magic Wand Demo 2021 by Rahim Entezari
0:53
[MCL] BlindApp: Free path for visually impaired people by Philip Samer,  2022
Embedded AI Lab @TUG
[MCL] BlindApp: Free path for visually impaired people by Philip Samer, 2022
2:32
[MCL] PlantApp: Garden plant identification and diagnostics by Alex Palmisano, 2022
Embedded AI Lab @TUG
[MCL] PlantApp: Garden plant identification and diagnostics by Alex Palmisano, 2022
1:16
[MCL] SpeechMood: Track your expressiveness by Maximilian Nothnagel, 2022
Embedded AI Lab @TUG
[MCL] SpeechMood: Track your expressiveness by Maximilian Nothnagel, 2022
0:53
[MCL] Activity Recognition by Julian Rudolf and Christopher Hinterer, 2022
Embedded AI Lab @TUG
[MCL] Activity Recognition by Julian Rudolf and Christopher Hinterer, 2022
3:07
[MCL] Activity Recognition by Peter Prodinger, 2020
Embedded AI Lab @TUG
[MCL] Activity Recognition by Peter Prodinger, 2020
0:43
[MCL] Activity Recognition by Jakob Soukup, 2022
Embedded AI Lab @TUG
[MCL] Activity Recognition by Jakob Soukup, 2022
0:48
[MCL] Activity Recognition by Dietmar Malli, 2020
Embedded AI Lab @TUG
[MCL] Activity Recognition by Dietmar Malli, 2020
0:47
[Embedded ML] Magic Wand Demo 2022 by David Mihola
Embedded AI Lab @TUG
[Embedded ML] Magic Wand Demo 2022 by David Mihola
2:44