Beyond Correlation: A New Approach to Graph Classification
Researchers have developed a novel graph neural network that uses causal reasoning to improve accuracy and reliability in identifying key features within complex graph structures.
Science
Words Matter: How Report Sentiment Impacts Bank Profits
New research reveals a direct causal link between the language used in financial reports and a bank’s bottom line.
Weather AI: Closing the Disaster Warning Gap in Africa
A new production-grade artificial intelligence system is providing national-scale weather forecasting and early warnings across South Africa, dramatically lowering the cost of disaster preparedness.
When Conversations Go Wrong: Teaching AI Agents to Recover
A new approach empowers conversational AI to identify and correct errors during complex, multi-turn interactions.
Unmasking Online Identities: The Rise of Automated Deanonymization
New research reveals that artificial intelligence can now automatically link pseudonymous online accounts to real-world identities with alarming accuracy.
Rating Sustainability: Can AI Build Trust?
A new framework leverages the power of artificial intelligence and human expertise to assess the reliability of corporate sustainability ratings.
Learning from the Past: AI Improves with Memory-Enhanced Reinforcement
A new framework empowers artificial intelligence to learn more efficiently by retrospectively analyzing past experiences and assigning credit where it’s due.
Mapping Dementia: A New Network Approach to Early Diagnosis
![PaIRWaL encodes random walks on similarity networks derived from gyral folding to produce invariant sequences, subsequently aggregated to facilitate graph-level classification-a process fundamentally reliant on the topological properties of the underlying data and yielding representations insensitive to node permutations [latex] \mathbb{R}^n [/latex].](https://arxiv.org/html/2602.17557v1/x1.png)
Researchers are leveraging the unique folding patterns of the brain to build more accurate diagnostic tools for Alzheimer’s and Lewy body dementia.
When Algorithms Get It Wrong: Navigating Imperfect Models
New research reveals how Bayesian algorithms like Thompson Sampling behave when faced with inaccurate models, offering insights into their robustness and potential pitfalls.
Riding the Chaos: Accurately Predicting System Responses to Complex Forces
![The system’s state space trajectories, captured at intervals of approximately 10 seconds, demonstrate the predictive capacity of a Gaussian State Space model (red) and an autoencoder-LSTM network (green) against the backdrop of a periodically forced system (cyan) and its true response (black) as projected onto different phase space coordinates-specifically, [latex]x_{10}, \dot{x}_{10}, x_{11}[/latex] and [latex]x_1, \dot{x}_1, x_{20}[/latex]-revealing the models’ ability to approximate system behavior across varying states.](https://arxiv.org/html/2602.16848v1/x16.png)
A new computational approach offers a faster and more precise method for determining the stable states of mechanical systems under unpredictable external influences.