9 citations · 16 across the 15 of their papers we have counts for
20 papers
COMPASS: Finding Where Reasoning Lives in Language Models
Pratyay Dutta, Kowshik Thopalli, Vivek Narayanaswamy
Explicitly eliciting reasoning substantially improves LLM performance. Existing approaches require a predefined characterization of reasoning, whether through CoT prompt design, co…
Unsupervised Latent Space Alignment with Hyperspherical Geodesic Matching
Cameron Ryan, Vivek Sivaraman Narayanaswamy, Kowshik Thopalli +1
Independently trained neural networks tend to encode the same data with similar latent geometries. These latent geometries are not directly compatible, yet they can be nearly the s…
Improving Robustness In Sparse Autoencoders via Masked Regularization
Vivek Narayanaswamy, Kowshik Thopalli, Bhavya Kailkhura +1
Sparse autoencoders (SAEs) are widely used in mechanistic interpretability to project LLM activations onto sparse latent spaces. However, sparsity alone is an imperfect proxy for i…
The Anatomy of Uncertainty in LLMs
Aditya Taparia, Ransalu Senanayake, Kowshik Thopalli +1
Understanding why a large language model (LLM) is uncertain about the response is important for their reliable deployment. Current approaches, which either provide a single uncerta…
Interpretable and Steerable Concept Bottleneck Sparse Autoencoders
Akshay Kulkarni, Tsui-Wei Weng, Vivek Narayanaswamy +3
Sparse autoencoders (SAEs) promise a unified approach for mechanistic interpretability, concept discovery, and model steering in LLMs and LVLMs. However, realizing this potential r…
Leveraging Registers in Vision Transformers for Robust Adaptation
Srikar Yellapragada, Kowshik Thopalli, Vivek Narayanaswamy +5
Vision Transformers (ViTs) have shown success across a variety of tasks due to their ability to capture global image representations. Recent studies have identified the existence o…
DECIDER: Leveraging Foundation Model Priors for Improved Model Failure Detection and Explanation
Rakshith Subramanyam, Kowshik Thopalli, Vivek Narayanaswamy +1
Reliably detecting when a deployed machine learning model is likely to fail on a given input is crucial for ensuring safe operation. In this work, we propose DECIDER (Debiasing Cla…
On the Use of Anchoring for Training Vision Models
Vivek Narayanaswamy, Kowshik Thopalli, Rushil Anirudh +3
Anchoring is a recent, architecture-agnostic principle for training deep neural networks that has been shown to significantly improve uncertainty estimation, calibration, and extra…
PAGER: A Framework for Failure Analysis of Deep Regression Models
Jayaraman J. Thiagarajan, Vivek Narayanaswamy, Puja Trivedi +1
Safe deployment of AI models requires proactive detection of failures to prevent costly errors. To this end, we study the important problem of detecting failures in deep regression…
An L2-Normalized Spatial Attention Network For Accurate And Fast Classification Of Brain Tumors In 2D T1-Weighted CE-MRI Images
Grace Billingsley, Julia Dietlmeier, Vivek Narayanaswamy +2
We propose an accurate and fast classification network for classification of brain tumors in MRI images that outperforms all lightweight methods investigated in terms of accuracy.…
Single Model Uncertainty Estimation via Stochastic Data Centering
Jayaraman J. Thiagarajan, Rushil Anirudh, Vivek Narayanaswamy +1
We are interested in estimating the uncertainties of deep neural networks, which play an important role in many scientific and engineering problems. In this paper, we present a str…
Know Your Space: Inlier and Outlier Construction for Calibrating Medical OOD Detectors
Vivek Narayanaswamy, Yamen Mubarka, Rushil Anirudh +3
We focus on the problem of producing well-calibrated out-of-distribution (OOD) detectors, in order to enable safe deployment of medical image classifiers. Motivated by the difficul…
Designing Counterfactual Generators using Deep Model Inversion
Jayaraman J. Thiagarajan, Vivek Narayanaswamy, Deepta Rajan +3
Explanation techniques that synthesize small, interpretable changes to a given image while producing desired changes in the model prediction have become popular for introspecting b…
On the Design of Deep Priors for Unsupervised Audio Restoration
Vivek Sivaraman Narayanaswamy, Jayaraman J. Thiagarajan, Andreas Spanias
Unsupervised deep learning methods for solving audio restoration problems extensively rely on carefully tailored neural architectures that carry strong inductive biases for definin…
Loss Estimators Improve Model Generalization
Vivek Narayanaswamy, Jayaraman J. Thiagarajan, Deepta Rajan +1
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 pre…
Using Deep Image Priors to Generate Counterfactual Explanations
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. Th…
Accurate and Robust Feature Importance Estimation under Distribution Shifts
Jayaraman J. Thiagarajan, Vivek Narayanaswamy, Rushil Anirudh +2
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 use…
Unsupervised Audio Source Separation using Generative Priors
Vivek Narayanaswamy, Jayaraman J. Thiagarajan, Rushil Anirudh +1
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 tim…
Audio Source Separation via Multi-Scale Learning with Dilated Dense U-Nets
Vivek Sivaraman Narayanaswamy, Sameeksha Katoch, Jayaraman J. Thiagarajan +2
Modern audio source separation techniques rely on optimizing sequence model architectures such as, 1D-CNNs, on mixture recordings to generalize well to unseen mixtures. Specificall…
Designing an Effective Metric Learning Pipeline for Speaker Diarization
Vivek Sivaraman Narayanaswamy, Jayaraman J. Thiagarajan, Huan Song +1
State-of-the-art speaker diarization systems utilize knowledge from external data, in the form of a pre-trained distance metric, to effectively determine relative speaker identitie…