Zobrazeno 1 - 10
of 11 577
pro vyhledávání: '"A Jeevan"'
We introduce a new metric to assess the quality of generated images that is more reliable, data-efficient, compute-efficient, and adaptable to new domains than the previous metrics, such as Fr\'echet Inception Distance (FID). The proposed metric is b
Externí odkaz:
http://arxiv.org/abs/2411.15584
Our study introduces ResNet-L2 (RL2), a novel metric for evaluating generative models and image quality in histopathology, addressing limitations of traditional metrics, such as Frechet inception distance (FID), when the data is scarce. RL2 leverages
Externí odkaz:
http://arxiv.org/abs/2411.01034
We propose two new evaluation metrics to assess realness of generated images based on normalizing flows: a simpler and efficient flow-based likelihood distance (FLD) and a more exact dual-flow based likelihood distance (D-FLD). Because normalizing fl
Externí odkaz:
http://arxiv.org/abs/2410.02004
Federated learning faces a critical challenge in balancing communication efficiency with rapid convergence, especially for second-order methods. While Newton-type algorithms achieve linear convergence in communication rounds, transmitting full Hessia
Externí odkaz:
http://arxiv.org/abs/2409.15216
Current video summarization methods largely rely on transformer-based architectures, which, due to their quadratic complexity, require substantial computational resources. In this work, we address these inefficiencies by enhancing the Direct-to-Summa
Externí odkaz:
http://arxiv.org/abs/2409.14724
Recent advancements in single image super-resolution have been predominantly driven by token mixers and transformer architectures. WaveMixSR utilized the WaveMix architecture, employing a two-dimensional discrete wavelet transform for spatial token m
Externí odkaz:
http://arxiv.org/abs/2409.10582
Autor:
Chandra, Jeevan
We describe a one-parameter family of Euclidean wormhole solutions with the topology of a compact hyperbolic space times an interval in Einstein gravity minimally coupled to a massless scalar field in AdS$_{d+1}$ commonly referred to as Einstein-dila
Externí odkaz:
http://arxiv.org/abs/2407.15630
Autor:
Jeevan, Pranav, Sethi, Amit
In contemporary computer vision applications, particularly image classification, architectural backbones pre-trained on large datasets like ImageNet are commonly employed as feature extractors. Despite the widespread use of these pre-trained convolut
Externí odkaz:
http://arxiv.org/abs/2406.05612
We study the efficacy of Small Language Models (SLMs) in facilitating application usage through natural language interactions. Our focus here is on a particular internal application used in Microsoft for cloud supply chain fulfilment. Our experiments
Externí odkaz:
http://arxiv.org/abs/2405.20347
Autor:
Menache, Ishai1 (AUTHOR), Pathuri, Jeevan2 (AUTHOR), Simchi-Levi, David3 (AUTHOR), Linton, Tom4 (AUTHOR)
Publikováno v:
Harvard Business Review. Jan/Feb2025, Vol. 103 Issue 1, p86-95. 10p. 6 Color Photographs.
