Megh Shukla
EPFL ENAC IIC VITA
GC C1 392 (Bâtiment GC)
Station 18
1015 Lausanne
Web site: Web site: https://vita.epfl.ch/
Publications
Infoscience publications
TIC-TAC: A Framework for Improved Covariance Estimation in Deep Heteroscedastic Regression
Deep heteroscedastic regression involves jointly optimizing the mean and covariance of the predicted distribution using the negative log-likelihood. However, recent works show that this may result in sub-optimal convergence due to the challenges associated with covariance estimation. While the literature addresses this by proposing alternate formulations to mitigate the impact of the predicted covariance, we focus on improving this predicted covariance itself. We study two questions: (1) Does the predicted covariance truly capture the randomness of the predicted mean? (2) In the absence of supervision, how can we quantify the accuracy of covariance estimation? We address (1) with a Taylor Induced Covariance (TIC), which captures the randomness of the predicted mean by incorporating its gradient and curvature through the second order Taylor polynomial. Furthermore, we tackle (2) by introducing the Task Agnostic Correlations (TAC) metric, which combines the notion of correlations and absolute error to evaluate the covariance. We evaluate TIC-TAC across multiple experiments spanning synthetic and real-world datasets. Our results show that not only does TIC accurately learn the covariance, it additionally facilitates an improved convergence of negative log-likelihood. We make our code available at: https://github.com/vita-epfl/TIC-TAC
2024-07-21. 41st International Conference on Machine Learning (ICML) 2024, Vienna, Austria, July 21-27, 2024.VL4Pose: Active Learning Through Out-Of-Distribution Detection For Pose Estimation
Advances in computing have enabled widespread access to pose estimation, creating new sources of data streams. Unlike mock set-ups for data collection, tapping into these data streams through on-device active learning allows us to directly sample from the real world to improve the spread of the training distribution. However, on-device computing power is limited, implying that any candidate active learning algorithm should have a low compute footprint while also being reliable. Although multiple algorithms cater to pose estimation, they either use extensive compute to power state-of-the-art results or are not competitive in low-resource settings. We address this limitation with VL4Pose (Visual Likelihood For Pose Estimation), a first principles approach for active learning through out-of-distribution detection. We begin with a simple premise: pose estimators often predict incoherent poses for out-of-distribution samples. Hence, can we identify a distribution of poses the model has been trained on, to identify incoherent poses the model is unsure of? Our solution involves modelling the pose through a simple parametric Bayesian network trained via maximum likelihood estimation. Therefore, poses incurring a low likelihood within our framework are out-of-distribution samples making them suitable candidates for annotation. We also observe two useful side-outcomes: VL4Pose in-principle yields better uncertainty estimates by unifying joint and pose level ambiguity, as well as the unintentional but welcome ability of VL4Pose to perform pose refinement in limited scenarios. We perform qualitative and quantitative experiments on three datasets: MPII, LSP and ICVL, spanning human and hand pose estimation. Finally, we note that VL4Pose is simple, computationally inexpensive and competitive, making it suitable for challenging tasks such as on-device active learning.
2022-11-21. 33rd British Machine Vision Conference (BMVC 2022), London, UK, November 21-24, 2022.Selected publications
| Megh Shukla IEEE/CVF WACV 2022 |
Bayesian Uncertainty and Expected Gradient Length - Regression: Two Sides of the Same Coin? |
| Megh Shukla et al. IEEE/CVF CVPR-W 2021 |
A Mathematical Analysis of Learning Loss for Active Learning in Regression |
| Megh Shukla et al. IEEE ICASSP 2020 |
LEt-SNE: A Hybrid Approach to Data Embedding and Visualization of Satellite Imagery |