Zobrazeno 1 - 10
of 18
pro vyhledávání: '"Shtedritski, Aleksandar"'
Canonical surface mapping generalizes keypoint detection by assigning each pixel of an object to a corresponding point in a 3D template. Popularised by DensePose for the analysis of humans, authors have since attempted to apply the concept to more ca
Externí odkaz:
http://arxiv.org/abs/2407.18907
This paper investigates biases of Large Language Models (LLMs) through the lens of grammatical gender. Drawing inspiration from seminal works in psycholinguistics, particularly the study of gender's influence on language perception, we leverage multi
Externí odkaz:
http://arxiv.org/abs/2407.09704
Autor:
Franzmeyer, Tim, Shtedritski, Aleksandar, Albanie, Samuel, Torr, Philip, Henriques, João F., Foerster, Jakob N.
Benchmarks have been essential for driving progress in machine learning. A better understanding of LLM capabilities on real world tasks is vital for safe development. Designing adequate LLM benchmarks is challenging: Data from real-world tasks is har
Externí odkaz:
http://arxiv.org/abs/2406.03428
Autor:
Lála, Jakub, O'Donoghue, Odhran, Shtedritski, Aleksandar, Cox, Sam, Rodriques, Samuel G., White, Andrew D.
Large Language Models (LLMs) generalize well across language tasks, but suffer from hallucinations and uninterpretability, making it difficult to assess their accuracy without ground-truth. Retrieval-Augmented Generation (RAG) models have been propos
Externí odkaz:
http://arxiv.org/abs/2312.07559
Autor:
O'Donoghue, Odhran, Shtedritski, Aleksandar, Ginger, John, Abboud, Ralph, Ghareeb, Ali Essa, Booth, Justin, Rodriques, Samuel G
The ability to automatically generate accurate protocols for scientific experiments would represent a major step towards the automation of science. Large Language Models (LLMs) have impressive capabilities on a wide range of tasks, such as question a
Externí odkaz:
http://arxiv.org/abs/2310.10632
Autor:
Hall, Siobhan Mackenzie, Abrantes, Fernanda Gonçalves, Zhu, Hanwen, Sodunke, Grace, Shtedritski, Aleksandar, Kirk, Hannah Rose
We introduce VisoGender, a novel dataset for benchmarking gender bias in vision-language models. We focus on occupation-related biases within a hegemonic system of binary gender, inspired by Winograd and Winogender schemas, where each image is associ
Externí odkaz:
http://arxiv.org/abs/2306.12424
Autor:
Smith, Brandon, Farinha, Miguel, Hall, Siobhan Mackenzie, Kirk, Hannah Rose, Shtedritski, Aleksandar, Bain, Max
Vision-language models are growing in popularity and public visibility to generate, edit, and caption images at scale; but their outputs can perpetuate and amplify societal biases learned during pre-training on uncurated image-text pairs from the int
Externí odkaz:
http://arxiv.org/abs/2305.15407
Large-scale Vision-Language Models, such as CLIP, learn powerful image-text representations that have found numerous applications, from zero-shot classification to text-to-image generation. Despite that, their capabilities for solving novel discrimin
Externí odkaz:
http://arxiv.org/abs/2304.06712
Autor:
Berg, Hugo, Hall, Siobhan Mackenzie, Bhalgat, Yash, Yang, Wonsuk, Kirk, Hannah Rose, Shtedritski, Aleksandar, Bain, Max
Vision-language models can encode societal biases and stereotypes, but there are challenges to measuring and mitigating these multimodal harms due to lacking measurement robustness and feature degradation. To address these challenges, we investigate
Externí odkaz:
http://arxiv.org/abs/2203.11933
Autor:
Kirk, Hannah Rose, Jun, Yennie, Rauba, Paulius, Wachtel, Gal, Li, Ruining, Bai, Xingjian, Broestl, Noah, Doff-Sotta, Martin, Shtedritski, Aleksandar, Asano, Yuki M.
Hateful memes pose a unique challenge for current machine learning systems because their message is derived from both text- and visual-modalities. To this effect, Facebook released the Hateful Memes Challenge, a dataset of memes with pre-extracted te
Externí odkaz:
http://arxiv.org/abs/2107.04313