Skip to content

usdaed

  • Science
  • Who is Denis Avetissian?

Science

Smart Data for Smarter Robots: Distilling Knowledge from Vast Experience

23.11.2025 by qfx

A novel framework distills high-value synthetic datasets from extensive robot demonstration data via a three-stage pipeline: multimodal representation learning encodes raw data streams, a two-stage influence assessment engine-utilizing influence functions and contrastive verification with programmatically generated minimal counterexamples-quantifies sample importance, and influence-guided non-conformity filtering distills a synthetic coreset by training an adversarial network to maximize feature distribution coverage, as evidenced by t-SNE visualizations demonstrating successful replication of original high-value sample density.

A new framework efficiently compresses large vision-language-action datasets into smaller, more manageable sets for training robot learning models.

Categories Science

Spiking Neurons Tune In: AI for Cleaner Radio Signals

22.11.2025 by qfx

A unified spiking neural network, trained as a large model and subsequently fragmented, facilitates radio frequency interference detection across diverse neuromorphic platforms-including both snnTorch simulations and SynSense Xylo hardware-with latency encoding of spectrograms and deployment in a standardized NIR format enabling comparative power measurements.

Researchers have developed an end-to-end artificial intelligence pipeline leveraging spiking neural networks to effectively filter out unwanted radio interference and enhance the clarity of astronomical data.

Categories Science

Unmasking Hidden Data: A Deep Learning Approach to Steganography

22.11.2025 by qfx

Researchers have developed a novel deep learning model capable of both detecting and recovering concealed information embedded within images using the APVD steganographic technique.

Categories Science

Rewriting Reality: Graph Diffusion for Explainable AI

22.11.2025 by qfx

Existing approaches to graph manipulation either lack support for discrete inputs, struggle with computational scalability, or fail to guarantee solutions remain within the data manifold; however, this work introduces a method satisfying all three criteria through distillation of label information into a conditional discrete diffusion model and a generation pipeline leveraging gradient-based conditional estimation ($GDCE$).

A new framework leverages the power of graph diffusion models to generate realistic and actionable counterfactual explanations for graph-structured data.

Categories Science

Smarter Networks: AI’s Role in 6G Energy Efficiency

22.11.2025 by qfx

The evolving dynamics inherent in 6G systems significantly influence energy efficiency, demanding a holistic consideration of these interdependencies for optimized performance.

This review explores how artificial intelligence is being leveraged to create more sustainable and adaptable next-generation wireless communication systems.

Categories Science

Seeing Through the Smoke: AI’s Next Step in Wildfire Detection

21.11.2025 by qfx

The research demonstrates a method for augmenting smoke image datasets by leveraging real smoke imagery and corresponding masks to generate more realistic synthetic data, effectively minimizing the discrepancy between simulated and real-world conditions-a crucial step for robust computer vision applications.

New research explores how synthetic data and advanced machine learning can overcome the challenges of identifying wildfires in real-world imagery.

Categories Science

Unmasking Monero: A New Approach to Blockchain Forensics

21.11.2025 by qfx

The articulation of relational topology through ART-graphs provides a framework for understanding system evolution as a process of constrained decay, where interconnectedness defines the pathways of inevitable change.

Researchers have developed a novel graph-based framework to analyze transaction patterns in the privacy-focused cryptocurrency Monero, offering a path to detect illicit activity without breaking its core anonymity features.

Categories Science

Seeing Clearly: AI Sheds Light on Diabetic Retinopathy

21.11.2025 by qfx

A new deep learning framework combines advanced image analysis with explainable AI to improve both the accuracy and clinical understanding of diabetic retinopathy detection.

Categories Science

Decoding the Market: Information Theory’s Edge

21.11.2025 by qfx

Information-theoretic measures applied to S&P 500 returns from 2000 to 2025 reveal that Shannon entropy spikes during periods of market uncertainty-notably the 2008-2009 financial crisis and the COVID-19 pandemic-while Kullback-Leibler divergence identifies significant distributional shifts exceeding $ \mu + 2\sigma $ during crises, and normalized mutual information, typically below 0.05 during stable periods, surges during major market disruptions, collectively suggesting these measures can effectively characterize and quantify market regime changes.

A new framework leverages the principles of information theory to quantify market efficiency and refine financial risk management.

Categories Science

Accounting’s AI Future: Charting a New Research Course

21.11.2025 by qfx

As artificial intelligence reshapes the business world, this article provides a critical framework for accounting researchers to navigate the opportunities and challenges of this rapidly evolving landscape.

Categories Science
Older posts
Newer posts
← Previous Page1 … Page157 Page158 Page159 … Page168 Next →
© 2026 usdaed • Built with GeneratePress