Moshe Tshuva

Moshe Tshuva


Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

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

Representative Publications

View All Publications
2 October 2024 in Environmental Monitoring and Assessment
Verified
Journal Article
Open access
PDF file
JBJosef Berman
Josef Berman
BPBen Pinhasov
Ben Pinhasov
MTMoshe Tshuva
Moshe Tshuva
YAYehudit Aperstein
Yehudit Aperstein
This research introduces syN-BEATS, a novel ensemble deep learning model tailored for effective pollutant forecasting under conditions of limited data availability. Based on the N-BEATS architecture, syN-BEATS integrates various configurations with differing numbers of stacks and blocks, effectively combining weak and strong learning approaches. Our experiments show that syN-BEATS outperforms standard models, especially when using Bayesian optimization to fine-tune ensemble weights. The model consistently achieves low relative root mean square errors, proving its capacity for precise pollutant forecasting despite data constraints. A key aspect of this study is the use of data from only one meteorological and one air quality monitoring station per region, simulating environments with restricted monitoring capabilities. By applying this approach in regions with diverse climates and air quality levels, we thoroughly assess the model’s flexibility and resilience under different environmental conditions. The results highlight syN-BEATS’ ability to support the development of effective health alert systems that can detect specific airborne pollutants, even in areas with limited monitoring infrastructure. This advancement is crucial for enhancing environmental monitoring and public health management in under-resourced areas.

17 November 2025 in arXiv
Verified
Preprint
Open access
PDF file
NTNoam Tsfaty
Noam Tsfaty
AWAvishai Weizman
Avishai Weizman
LCLiav Cohen
Liav Cohen
MTMoshe Tshuva
Moshe Tshuva
YAYehudit Aperstein
Yehudit Aperstein
We address the challenge of detecting rare and diverse anomalies in surveillance videos using only video-level supervision. Our dual-backbone framework combines convolutional and transformer representations through top-k pooling, achieving 90.7% area under the curve (AUC) on the UCF-Crime dataset.

4 May 2026 in Applied Sciences
Verified
Journal Article
Open access
PDF file
SGSharon Gat
Sharon Gat
GAGal Aviram
Gal Aviram
AYAmir Yehudayoff
Amir Yehudayoff
HTHen Toledano
Hen Toledano
MTMoshe Tshuva
Moshe Tshuva
GPGabriel Polliack
Gabriel Polliack
Early detection of malignancy is imperative, yet existing diagnostic approaches struggle to identify small peripheral lesions. This study evaluated a novel imaging modality, heat diffusion analysis, to assess its ability to differentiate between malignant and normal lung tissue. Considering that lung cancer is the leading cause of cancer-related mortality worldwide, lung tumors were induced in mice in a preclinical ex vivo model to evaluate the proposed technology. The HTOScan System was used to analyze the thermal characteristics of 60 sites from excised lungs, including normal and abnormal regions. The algorithm classified pixels as high- or low-risk for malignancy. The HTOScan System demonstrated a high accuracy of 97%, with 94% sensitivity and 98% specificity compared to the gold standard of histopathology. The technology successfully differentiated abnormal from normal tissue ex vivo based on differences in thermal diffusivity. This proof-of-concept study suggests that combining heat diffusion imaging techniques with machine learning algorithms could enable the HTOScan System to identify malignant lesions accurately with high confidence. The technique shows promise as a real-time decision support tool for cancer detection, pending further in vivo validation. This novel functional-imaging approach could improve the identification of peripheral lesions and the guidance of biopsies during bronchoscopy.