Zobrazeno 1 - 10
of 875
pro vyhledávání: '"Amit, Ranjan"'
We introduce a precision polarization scheme for DNN inference that utilizes only very low and very high precision levels, assigning low precision to the majority of network weights and activations while reserving high precision paths for targeted er
Externí odkaz:
http://arxiv.org/abs/2411.05845
Autor:
Darabi, Nastaran, Jayasuriya, Dinithi, Naik, Devashri, Tulabandhula, Theja, Trivedi, Amit Ranjan
Adversarial attacks exploit vulnerabilities in a model's decision boundaries through small, carefully crafted perturbations that lead to significant mispredictions. In 3D vision, the high dimensionality and sparsity of data greatly expand the attack
Externí odkaz:
http://arxiv.org/abs/2409.12379
Autor:
Hashem, Maeesha Binte, Parpillon, Benjamin, Kumar, Divake, Jayasuria, Dinithi, Trivedi, Amit Ranjan
In this work, we propose "TimeFloats," an efficient train-in-memory architecture that performs 8-bit floating-point scalar product operations in the time domain. While building on the compute-in-memory paradigm's integrated storage and inferential co
Externí odkaz:
http://arxiv.org/abs/2409.00495
We present a novel statistical approach to incorporating uncertainty awareness in model-free distributional reinforcement learning involving quantile regression-based deep Q networks. The proposed algorithm, $\textit{Calibrated Evidential Quantile Re
Externí odkaz:
http://arxiv.org/abs/2402.07107
Autor:
Darabi, Nastaran, Shukla, Priyesh, Jayasuriya, Dinithi, Kumar, Divake, Stutts, Alex C., Trivedi, Amit Ranjan
This paper addresses the challenging problem of energy-efficient and uncertainty-aware pose estimation in insect-scale drones, which is crucial for tasks such as surveillance in constricted spaces and for enabling non-intrusive spatial intelligence i
Externí odkaz:
http://arxiv.org/abs/2401.17481
Autor:
Parente, Domenico, Darabi, Nastaran, Stutts, Alex C., Tulabandhula, Theja, Trivedi, Amit Ranjan
This paper introduces a lightweight uncertainty estimator capable of predicting multimodal (disjoint) uncertainty bounds by integrating conformal prediction with a deep-learning regressor. We specifically discuss its application for visual odometry (
Externí odkaz:
http://arxiv.org/abs/2309.11018
In the expanding landscape of AI-enabled robotics, robust quantification of predictive uncertainties is of great importance. Three-dimensional (3D) object detection, a critical robotics operation, has seen significant advancements; however, the major
Externí odkaz:
http://arxiv.org/abs/2309.09593
Autor:
Darabi, Nastaran, Hashem, Maeesha Binte, Pan, Hongyi, Cetin, Ahmet, Gomes, Wilfred, Trivedi, Amit Ranjan
The edge processing of deep neural networks (DNNs) is becoming increasingly important due to its ability to extract valuable information directly at the data source to minimize latency and energy consumption. Frequency-domain model compression, such
Externí odkaz:
http://arxiv.org/abs/2309.01771
Autor:
Giacomini, Davide, Hashem, Maeesha Binte, Suarez, Jeremiah, Bhunia, Swarup, Trivedi, Amit Ranjan
The rapid advancement of deep neural networks has significantly improved various tasks, such as image and speech recognition. However, as the complexity of these models increases, so does the computational cost and the number of parameters, making it
Externí odkaz:
http://arxiv.org/abs/2307.07631
Autor:
Nasrin, Shamma, Hashem, Maeesha Binte, Darabi, Nastaran, Parpillon, Benjamin, Fahim, Farah, Gomes, Wilfred, Trivedi, Amit Ranjan
This work discusses memory-immersed collaborative digitization among compute-in-memory (CiM) arrays to minimize the area overheads of a conventional analog-to-digital converter (ADC) for deep learning inference. Thereby, using the proposed scheme, si
Externí odkaz:
http://arxiv.org/abs/2307.03863