10 papers · 1 filter
Monitoring latent reasoning in looped language models
Pawel Olszowiec, Paulina Tomaszewska, Maciej Żelaszczyk +2
Looped language models (LMs) reuse one block of layers several times per token generation, promising more latent computation at no additional storage cost. Despite their rising pop…
Investigating Query-Insensitive Behavior in Spatio-Temporal Video Grounding
Eryk Kołodziejczyk, Alberto Presta, Karol Szurkowski +1
Spatio-temporal video grounding (STVG) aims to localize objects or events described by natural language queries in both space and time. Existing STVG models are typically trained a…
Pocket-STVG: lightweight architecture for Spatio-Temporal Video Grounding
Alberto Presta, Michal Byra, Grzegorz Stefański +3
Spatio-Temporal Video Grounding (STVG) aims to localize the spatio-temporal tube in a video corresponding to a natural language query. While recent methods achieve strong performan…
Same path, different walk: a mechanistic comparison of looped and stacked transformer encoders on 12-lead ECG
Pawel Olszowiec, Michal Byra, Grzegorz Gruszczynski +2
Recurrent Transformers reusing their weights rather than stacking distinct layers are becoming widely adopted due to their parameter efficiency [1,2,3]. However, the exact repr…
Zero-shot video highlight detection based on text descriptions and synthetic images
Michal Byra, Alberto Presta, Grzegorz Stefanski +1
Detecting video highlights, the most informative or engaging moments in a video, is important for applications such as video summarization and content recommendation. We propose a…
Training Crossroads for Recurrent Vision Transformers: Recurrence, Neural ODEs, and Deep Supervision
Grzegorz Gruszczynski, Pawel Olszowiec, Michal Byra +2
Vision Transformers (ViTs) achieve strong image-recognition performance, but their parameter count grows linearly with depth when each block is independently parameterized. Single-…
Weight-Space Mixture-of-Experts for Implicit Neural Representation Classification
Stanislaw Janik, Michal Byra
Implicit Neural Representations (INRs) encode signals as the weights of a coordinate-based neural network and have recently been proposed as an alternative domain for downstream le…
t-gems: text-guided exit modules for decreasing clip image encoder
Alberto Presta, Grzegorz Stefanski, Michal Byra +1
Multimodal deep neural networks enhance deep comprehension by integrating diverse data modalities. Data from different modalities are typically projected into a shared latent space…
Investigating Single-Block Recurrence in Vision Transformers for Image Recognition
Michal Byra, Pawel Olszowiec, Grzegorz Stefanski +2
Vision Transformers (ViTs) implement depth by stacking independently parameterized blocks, but it remains unclear how much of this parameterization is necessary and how much can be…
Routing the Lottery: Adaptive Subnetworks for Heterogeneous Data
Grzegorz Stefanski, Alberto Presta, Michal Byra
In pruning, the Lottery Ticket Hypothesis posits that large networks contain sparse subnetworks, or winning tickets, that can be trained in isolation to match the performance of th…