Ethan Fetaya

Prof. Dr. Ethan Fetaya


Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
25 July 2022 in Sensors
Verified
Journal Article
Open access
PDF file
Dan LeviDan Levi
Dan Levi
LGLiran Gispan
Liran Gispan
NGNiv Giladi
Niv Giladi
EFEthan Fetaya
Ethan Fetaya
Predicting not only the target but also an accurate measure of uncertainty is important for many machine learning applications, and in particular, safety-critical ones. In this work, we study the calibration of uncertainty prediction for regression tasks which often arise in real-world systems. We show that the existing definition for the calibration of regression uncertainty has severe limitations in distinguishing informative from non-informative uncertainty predictions. We propose a new definition that escapes this caveat and an evaluation method using a simple histogram-based approach. Our method clusters examples with similar uncertainty prediction and compares the prediction with the empirical uncertainty on these examples. We also propose a simple, scaling-based calibration method that preforms as well as much more complex ones. We show results on both a synthetic, controlled problem and on the object detection bounding-box regression task using the COCO and KITTI datasets.

17 March 2023 in Science
Verified
Journal Article
Closed access
SYShaofan Yuan
Shaofan Yuan
CMChao Ma
Chao Ma
EFEthan Fetaya
Ethan Fetaya
TMThomas Mueller
Thomas Mueller
DNDoron Naveh
Doron Naveh
FXFengnian Xia
Fengnian Xia
Geometry, an ancient yet vibrant branch of mathematics, has important and far-reaching impacts on various disciplines such as art, science, and engineering. Here, we introduce an emerging concept dubbed “geometric deep optical sensing” that is based on a number of recent demonstrations in advanced optical sensing and imaging, in which a reconfigurable sensor (or an array thereof) can directly decipher the rich information of an unknown incident light beam, including its intensity, spectrum, polarization, spatial features, and possibly angular momentum. We present the physical, mathematical, and engineering foundations of this concept, with particular emphases on the roles of classical and quantum geometry and deep neural networks. Furthermore, we discuss the new opportunities that this emerging scheme can enable and the challenges associated with future developments.

15 September 2020 in Proceedings of the National Academy of Sciences
Verified
Journal Article
Closed access
PDF file
EFEthan Fetaya
Ethan Fetaya
YLYonatan Lifshitz
Yonatan Lifshitz
EAElad Aaron
Elad Aaron
SGShai Gordin
Shai Gordin
Significance: The documentary sources for the political, economic, and social history of ancient Mesopotamia constitute hundreds of thousands of clay tablets inscribed in the cuneiform script. Most tablets are damaged, leaving gaps in the texts written on them, and the missing portions must be restored by experts. This paper uses available digitized texts for training advanced machine-learning algorithms to restore daily economic and administrative documents from the Persian empire (sixth to fourth centuries BCE). As the amount of digitized texts grows, the model can be trained to restore damaged texts belonging to other genres, such as scientific or literary texts. Therefore, this is a first step for a large-scale reconstruction of a lost ancient heritage.

Verified
Journal Article
Open access
PDF file
EBEyal Betzalel
Eyal Betzalel
CPCoby Penso
Coby Penso
EFEthan Fetaya
Ethan Fetaya
Generative models such as generative adversarial networks, diffusion models, and variational auto-encoders have become prevalent in recent years. While it is true that these models have shown remarkable results, evaluating their performance is challenging. This issue is of vital importance to push research forward and identify meaningful gains from random noise. Currently, heuristic metrics such as the inception score (IS) and Fréchet inception distance (FID) are the most common evaluation metrics, but what they measure is not entirely clear. Additionally, there are questions regarding how meaningful their score actually is. In this work, we propose a novel evaluation protocol for likelihood-based generative models, based on generating a high-quality synthetic dataset on which we can estimate classical metrics for comparison. This new scheme harnesses the advantages of knowing the underlying likelihood values of the data by measuring the divergence between the model-generated data and the synthetic dataset. Our study shows that while FID and IS correlate with several f-divergences, their ranking of close models can vary considerably, making them problematic when used for fine-grained comparison. We further use this experimental setting to study which evaluation metric best correlates with our probabilistic metrics.

6 October 2023 in IEEE Signal Processing Letters
Verified
Journal Article
Closed access
IAIdan Achituve
Idan Achituve
Wenbo WangWenbo Wang
Wenbo Wang
EFEthan Fetaya
Ethan Fetaya
ALAmir Leshem
Amir Leshem
Vertically distributed learning exploits the local features collected by multiple learning workers to form a better global model. However, data exchange between the workers and the model aggregator for parameter training incurs a heavy communication burden, especially when the learning system is built upon capacity-constrained wireless networks. In this paper, we propose a novel hierarchical distributed learning framework, where each worker separately learns a low-dimensional embedding of their local observed data. Then, they perform communication-efficient distributed max-pooling to efficiently transmit the synthesized input to the aggregator. For data exchange over a shared wireless channel, we propose an opportunistic carrier sensing-based protocol to implement the max-pooling of the output of all the workers. Our simulation experiments show that the proposed learning framework is able to achieve almost the same model accuracy as the learning model using the concatenation of all the raw outputs from the learning workers while significantly reducing the communication load.

15 March 2025 in Scientific Data
Verified
Journal Article
Open access
PDF file
JEJonathan Eby
Jonathan Eby
MBMoshe Beutel
Moshe Beutel
DKDavid Koivisto
David Koivisto
IAIdan Achituve
Idan Achituve
EFEthan Fetaya
Ethan Fetaya
JZJosé Zariffa
José Zariffa
Neurotechnological interfaces have the potential to create new forms of human-machine interactions, by allowing devices to interact directly with neurological signals instead of via intermediates such as keystrokes. Surface electromyography (sEMG) has been used extensively in myoelectric control systems, which use bioelectric activity recorded from muscles during contractions to classify actions. This technology has been used primarily for rehabilitation applications. In order to support the development of myoelectric interfaces for a broader range of human-machine interactions, we present an sEMG dataset obtained during key presses in a typing task. This fine-grained classification dataset consists of 16-channel bilateral sEMG recordings and key logs, collected from 19 individuals in two sessions on different days. We report baseline results on intra-session, inter-session and inter-subject evaluations. Our baseline results show that within-session accuracy is relatively high, even with simple learning models. However, the results on between-session and between-participant are much lower, showing that generalizing between sessions and individuals is an open challenge.