Science – Page 54

Seeing Through the Synthetic: Unmasking AI-Generated Video

31.01.2026 by qfx

The analysis reveals a fundamental distinction between artificially generated content, characterized by predictable, structured residuals stemming from a fixed basis function, and recordings of the physical world, which inherently exhibit heterogeneous, unstructured noise-a contrast highlighting the limitations of even advanced synthesis techniques in fully replicating natural complexity.

A new forensic approach leverages subtle inconsistencies in AI-created videos to reliably distinguish them from authentic footage.

Smarter Energy Data: One Model to Rule Them All

31.01.2026 by qfx

The study demonstrates that the proposed SmartMeterFM achieves superior 16x super-resolution performance compared to linear interpolation and ProfileSR, effectively reconstructing high-resolution details from low-resolution inputs-a result indicative of the algorithm’s capacity to model complex relationships within image data and surpass the limitations of simpler methods.

Researchers have developed a new approach to generating and refining smart meter data, offering a single model capable of handling multiple critical tasks.

Spotting the Spin: Better AI for Detecting Greenwashing

31.01.2026 by qfx

$The study demonstrates that parameter-efficient adaptation techniques, such as LoRA, and full fine-tuning of T5 models exhibit comparable performance across different data folds, as measured by [latex]|\Delta|[/latex], though variations arise based on specific configurations and underlying model architectures.$

New research details a method for improving the ability of artificial intelligence to reliably identify misleading claims about sustainability in corporate reports.

Smarter Data Sampling for Faster, More Accurate Machine Learning

31.01.2026 by qfx

$A coupled framework adaptively samples scalar fields by leveraging a Gaussian process surrogate-which estimates means [latex]\mu_{GP}[/latex] and variances [latex]\sigma_{GP}[/latex] from inputs [latex]\bm{\xi}[/latex]-and a field model that integrates scalar quantities to approximate means [latex]\mu_{GNN}[/latex] and variances [latex]\sigma_{GNN}[/latex], with the subsequent misfit and epistemic uncertainties driving an iterative infill criterion to refine sampling points and update both surrogates.$

A new strategy combines Gaussian processes and graph neural networks to dramatically improve the efficiency of building predictive models for complex field data.

Unveiling Stellar Birthwinds with Artificial Intelligence

31.01.2026 by qfx

$The study demonstrates that machine learning models-specifically ResNet50, ResNet152, ViT\_B\_16, and ViT\_L\_16-trained on deliberately incomplete datasets excluding specific protostellar masses and inclination angles, predictably falter when tasked with predicting those very excluded parameters, as evidenced by the divergence between ground-truth values and model predictions and the resulting skewed probability distribution functions of prediction errors.$

A new deep learning framework offers a powerful method for characterizing the complex outflows from young, massive stars.

Building Agents That Teach Themselves

31.01.2026 by qfx

A tool-chain pipeline streamlines the synthesis of trajectories, enabling efficient and systematic path planning.

A new framework automatically generates training data and environments, enabling language models to master complex, multi-step tasks through self-play.

Flow Networks Get a Boost from Smart Sampling

31.01.2026 by qfx

As the ratio of cardinality to non-cardinality and number of trajectories decreases, a submodular upper bound demonstrably expands coverage across terminating states by orders of magnitude compared to a classical Generalized Function Network, with scenarios exceeding the [latex]1:1:1[/latex] ratio between query and coverage indicating the value of this approach.

A new technique leverages the structure of rewards to guide the exploration of generative flow networks, leading to more efficient and effective solution discovery.

Guarding Federated Learning Against Bad Actors

31.01.2026 by qfx

Following a period of consistent performance across all clients, deviations in reported loss begin to emerge once a malicious client initiates manipulation, indicating a disruption in the expected convergence behavior of the distributed system and highlighting the potential for adversarial influence on global model training.

A new approach monitors how quickly clients learn to identify and neutralize those manipulating the system.

The Untapped Potential of Public Model Hubs

31.01.2026 by qfx

New research reveals that readily available model repositories contain surprisingly effective, yet overlooked, models that can significantly boost performance.

Planning’s Peril: Why Model-Based RL Struggles to Find the Right Path

31.01.2026 by qfx

Despite the promise of efficient learning, model-based reinforcement learning often falters due to unexpected challenges in its planning process.