Zobrazeno 1 - 10
of 90
pro vyhledávání: '"Oh, Seong Joon"'
Retrieval-augmented generation (RAG) mitigates many problems of fully parametric language models, such as temporal degradation, hallucinations, and lack of grounding. In RAG, the model's knowledge can be updated from documents provided in context. Th
Externí odkaz:
http://arxiv.org/abs/2404.16032
It has recently been conjectured that neural network solution sets reachable via stochastic gradient descent (SGD) are convex, considering permutation invariances (Entezari et al., 2022). This means that a linear path can connect two independent solu
Externí odkaz:
http://arxiv.org/abs/2403.07968
As large language models (LLMs) are increasingly deployed in user-facing applications, building trust and maintaining safety by accurately quantifying a model's confidence in its prediction becomes even more important. However, finding effective ways
Externí odkaz:
http://arxiv.org/abs/2403.05973
Uncertainty quantification, once a singular task, has evolved into a spectrum of tasks, including abstained prediction, out-of-distribution detection, and aleatoric uncertainty quantification. The latest goal is disentanglement: the construction of m
Externí odkaz:
http://arxiv.org/abs/2402.19460
Accurate uncertainty estimation is vital to trustworthy machine learning, yet uncertainties typically have to be learned for each task anew. This work introduces the first pretrained uncertainty modules for vision models. Similar to standard pretrain
Externí odkaz:
http://arxiv.org/abs/2402.16569
Large Language Model (LLM) services and models often come with legal rules on who can use them and how they must use them. Assessing the compliance of the released LLMs is crucial, as these rules protect the interests of the LLM contributor and preve
Externí odkaz:
http://arxiv.org/abs/2402.12991
Autor:
Scimeca, Luca, Rubinstein, Alexander, Teney, Damien, Oh, Seong Joon, Nicolicioiu, Armand Mihai, Bengio, Yoshua
Spurious correlations in the data, where multiple cues are predictive of the target labels, often lead to a phenomenon known as shortcut learning, where a model relies on erroneous, easy-to-learn cues while ignoring reliable ones. In this work, we pr
Externí odkaz:
http://arxiv.org/abs/2311.16176
Explainable AI (XAI) aims to provide insight into opaque model reasoning to humans and as such is an interdisciplinary field by nature. In this paper, we interviewed 10 practitioners to understand the possible usability of training data attribution (
Externí odkaz:
http://arxiv.org/abs/2310.20477
As machine learning technology gets applied to actual products and solutions, new challenges have emerged. Models unexpectedly fail to generalize to small changes in the distribution, tend to be confident on novel data they have never seen, or cannot
Externí odkaz:
http://arxiv.org/abs/2310.08215
Representation learning has significantly driven the field to develop pretrained models that can act as a valuable starting point when transferring to new datasets. With the rising demand for reliable machine learning and uncertainty quantification,
Externí odkaz:
http://arxiv.org/abs/2307.03810