Zobrazeno 1 - 10
of 34
pro vyhledávání: '"Hintersdorf, Dominik"'
Autor:
Hintersdorf, Dominik, Struppek, Lukas, Kersting, Kristian, Dziedzic, Adam, Boenisch, Franziska
Diffusion models (DMs) produce very detailed and high-quality images. Their power results from extensive training on large amounts of data, usually scraped from the internet without proper attribution or consent from content creators. Unfortunately,
Externí odkaz:
http://arxiv.org/abs/2406.02366
The proliferation of large language models (LLMs) has sparked widespread and general interest due to their strong language generation capabilities, offering great potential for both industry and research. While previous research delved into the secur
Externí odkaz:
http://arxiv.org/abs/2402.09132
The proliferation of large AI models trained on uncurated, often sensitive web-scraped data has raised significant privacy concerns. One of the concerns is that adversaries can extract information about the training data using privacy attacks. Unfort
Externí odkaz:
http://arxiv.org/abs/2310.08320
Label smoothing -- using softened labels instead of hard ones -- is a widely adopted regularization method for deep learning, showing diverse benefits such as enhanced generalization and calibration. Its implications for preserving model privacy, how
Externí odkaz:
http://arxiv.org/abs/2310.06549
Autor:
Struppek, Lukas, Hentschel, Martin B., Poth, Clifton, Hintersdorf, Dominik, Kersting, Kristian
Backdoor attacks pose a serious security threat for training neural networks as they surreptitiously introduce hidden functionalities into a model. Such backdoors remain silent during inference on clean inputs, evading detection due to inconspicuous
Externí odkaz:
http://arxiv.org/abs/2310.06372
The field of artificial intelligence (AI) has experienced remarkable progress in recent years, driven by the widespread adoption of open-source machine learning models in both research and industry. Considering the resource-intensive nature of traini
Externí odkaz:
http://arxiv.org/abs/2308.09490
Autor:
Struppek, Lukas, Hintersdorf, Dominik, Friedrich, Felix, Brack, Manuel, Schramowski, Patrick, Kersting, Kristian
Neural network-based image classifiers are powerful tools for computer vision tasks, but they inadvertently reveal sensitive attribute information about their classes, raising concerns about their privacy. To investigate this privacy leakage, we intr
Externí odkaz:
http://arxiv.org/abs/2303.09289
Autor:
Friedrich, Felix, Brack, Manuel, Struppek, Lukas, Hintersdorf, Dominik, Schramowski, Patrick, Luccioni, Sasha, Kersting, Kristian
Generative AI models have recently achieved astonishing results in quality and are consequently employed in a fast-growing number of applications. However, since they are highly data-driven, relying on billion-sized datasets randomly scraped from the
Externí odkaz:
http://arxiv.org/abs/2302.10893
Autor:
Brack, Manuel, Friedrich, Felix, Hintersdorf, Dominik, Struppek, Lukas, Schramowski, Patrick, Kersting, Kristian
Text-to-image diffusion models have recently received a lot of interest for their astonishing ability to produce high-fidelity images from text only. However, achieving one-shot generation that aligns with the user's intent is nearly impossible, yet
Externí odkaz:
http://arxiv.org/abs/2301.12247
Autor:
Brack, Manuel, Schramowski, Patrick, Friedrich, Felix, Hintersdorf, Dominik, Kersting, Kristian
Large, text-conditioned generative diffusion models have recently gained a lot of attention for their impressive performance in generating high-fidelity images from text alone. However, achieving high-quality results is almost unfeasible in a one-sho
Externí odkaz:
http://arxiv.org/abs/2212.06013