no code implementations • 20 Sep 2023 • Jayaraman J. Thiagarajan, Vivek Narayanaswamy, Puja Trivedi, Rushil Anirudh
In this paper, we propose PAGER (Principled Analysis of Generalization Errors in Regressors), a framework to systematically detect and characterize failures in deep regression models.
1 code implementation • 1 Aug 2023 • Grace Billingsley, Julia Dietlmeier, Vivek Narayanaswamy, Andreas Spanias, Noel E. OConnor
We compare our results against the state-of-the-art on this dataset and show that by integrating l2-normalized spatial attention into a baseline network we achieve a performance gain of 1. 79 percentage points.
1 code implementation • 14 Jul 2022 • Jayaraman J. Thiagarajan, Rushil Anirudh, Vivek Narayanaswamy, Peer-Timo Bremer
We are interested in estimating the uncertainties of deep neural networks, which play an important role in many scientific and engineering problems.
no code implementations • 12 Jul 2022 • Vivek Narayanaswamy, Yamen Mubarka, Rushil Anirudh, Deepta Rajan, Andreas Spanias, Jayaraman J. Thiagarajan
We focus on the problem of producing well-calibrated out-of-distribution (OOD) detectors, in order to enable safe deployment of medical image classifiers.
no code implementations • NeurIPS 2021 • Jayaraman J. Thiagarajan, Vivek Narayanaswamy, Deepta Rajan, Jason Liang, Akshay Chaudhari, Andreas Spanias
Explanation techniques that synthesize small, interpretable changes to a given image while producing desired changes in the model prediction have become popular for introspecting black-box models.
no code implementations • 5 Mar 2021 • Vivek Narayanaswamy, Jayaraman J. Thiagarajan, Deepta Rajan, Andreas Spanias
With increased interest in adopting AI methods for clinical diagnosis, a vital step towards safe deployment of such tools is to ensure that the models not only produce accurate predictions but also do not generalize to data regimes where the training data provide no meaningful evidence.
no code implementations • 22 Oct 2020 • Vivek Narayanaswamy, Jayaraman J. Thiagarajan, Andreas Spanias
Through the use of carefully tailored convolutional neural network architectures, a deep image prior (DIP) can be used to obtain pre-images from latent representation encodings.
no code implementations • 30 Sep 2020 • Jayaraman J. Thiagarajan, Vivek Narayanaswamy, Rushil Anirudh, Peer-Timo Bremer, Andreas Spanias
With increasing reliance on the outcomes of black-box models in critical applications, post-hoc explainability tools that do not require access to the model internals are often used to enable humans understand and trust these models.
1 code implementation • 28 May 2020 • Vivek Narayanaswamy, Jayaraman J. Thiagarajan, Rushil Anirudh, Andreas Spanias
State-of-the-art under-determined audio source separation systems rely on supervised end-end training of carefully tailored neural network architectures operating either in the time or the spectral domain.