Zobrazeno 1 - 10
of 918
pro vyhledávání: '"Pantazopoulos P"'
Autor:
Nikandrou, Malvina, Pantazopoulos, Georgios, Vitsakis, Nikolas, Konstas, Ioannis, Suglia, Alessandro
As Vision and Language models (VLMs) become accessible across the globe, it is important that they demonstrate cultural knowledge. In this paper, we introduce CROPE, a visual question answering benchmark designed to probe the knowledge of culture-spe
Externí odkaz:
http://arxiv.org/abs/2410.15453
Autor:
Pantazopoulos, Georgios, Nikandrou, Malvina, Suglia, Alessandro, Lemon, Oliver, Eshghi, Arash
This study explores replacing Transformers in Visual Language Models (VLMs) with Mamba, a recent structured state space model (SSM) that demonstrates promising performance in sequence modeling. We test models up to 3B parameters under controlled cond
Externí odkaz:
http://arxiv.org/abs/2409.05395
Continual learning focuses on incrementally training a model on a sequence of tasks with the aim of learning new tasks while minimizing performance drop on previous tasks. Existing approaches at the intersection of Continual Learning and Visual Quest
Externí odkaz:
http://arxiv.org/abs/2406.19297
Publikováno v:
Nature Communications 15, 4000 (2024)
Magnetic order typically emerges due to the short-range exchange interaction between the constituent electronic spins. Recent discoveries have found a crucial role for spin-phonon coupling in various phenomena from optical ultrafast magnetization swi
Externí odkaz:
http://arxiv.org/abs/2405.09246
Augmenting Large Language Models (LLMs) with image-understanding capabilities has resulted in a boom of high-performing Vision-Language models (VLMs). While studying the alignment of LLMs to human values has received widespread attention, the safety
Externí odkaz:
http://arxiv.org/abs/2405.04403
An effective method for combining frozen large language models (LLM) and visual encoders involves a resampler module that creates a `visual prompt' which is provided to the LLM, along with the textual prompt. While this approach has enabled impressiv
Externí odkaz:
http://arxiv.org/abs/2404.13594
Autor:
Pantazopoulos, Georgios, Nikandrou, Malvina, Parekh, Amit, Hemanthage, Bhathiya, Eshghi, Arash, Konstas, Ioannis, Rieser, Verena, Lemon, Oliver, Suglia, Alessandro
Interactive and embodied tasks pose at least two fundamental challenges to existing Vision & Language (VL) models, including 1) grounding language in trajectories of actions and observations, and 2) referential disambiguation. To tackle these challen
Externí odkaz:
http://arxiv.org/abs/2311.04067
Publikováno v:
Phys. Rev. B 109, L201408 (2024)
The use of cavity quantum electrodynamical effects, i.e., of vacuum electromagnetic fields, to modify material properties in cavities has rapidly gained popularity and interest in the last few years. However, there is still a scarcity of general resu
Externí odkaz:
http://arxiv.org/abs/2311.00183
Publikováno v:
Nature Communications, Vol 15, Iss 1, Pp 1-7 (2024)
Abstract Magnetic order typically emerges due to the short-range exchange interaction between the constituent electronic spins. Recent discoveries have found a crucial role for spin-phonon coupling in various phenomena from optical ultrafast magnetiz
Externí odkaz:
https://doaj.org/article/78f60ae9004547669fb86aca98f5bd00
Autor:
Christopher A. Martinez, Harry Pantazopoulos, Barbara Gisabella, Emily T. Stephens, Jacob Garteiser, Alberto Del Arco
Publikováno v:
Scientific Reports, Vol 14, Iss 1, Pp 1-13 (2024)
Abstract Repeated stress can predispose to substance abuse. However, behavioral and neurobiological adaptations that link stress to substance abuse remain unclear. This study investigates whether intermittent social defeat (ISD), a stress protocol th
Externí odkaz:
https://doaj.org/article/effde1a9be9a4b778918fd4c9e7c668a