Mapping Anomalies: A New Vision for Long-Term Trajectory Analysis
![TITAnD encodes the chaotic whispers of movement - whether dense GPS streams or sparse stay-points - into a unified hyperspectral trajectory image, where each pixel doesn’t simply mark location, but embodies a confluence of space, semantics, time, and the subtle poetry of motion [latex] day × time [/latex].](https://arxiv.org/html/2603.25255v1/imgs/intro-v5.png)
Researchers have developed a novel method for detecting unusual patterns in months-long GPS data by transforming movement into visual representations.
qfx
Predicting the Future: A New Approach to Irregular Time Series Forecasting
Researchers have developed a deep learning model that accurately forecasts complex systems even when data is missing or unevenly spaced in time.
Learning to Route: AI Surpasses Heuristics for the Traveling Salesman Problem
![The system calculates node embeddings from graph coordinates, then utilizes a causal transformer decoder-informed by both node transitions and predicted relational trajectory graphs [latex]\hat{R}_{t}[/latex]-to forecast subsequent relational trajectory graphs [latex]\tilde{R}_{t}[/latex], demonstrating a method for relational reasoning within dynamic graph structures.](https://arxiv.org/html/2603.25241v1/architecture.png)
A new approach using offline reinforcement learning trains an AI to find better routes than traditional methods, even without live experimentation.
Hidden Signals: Backdooring Deep Learning for Wireless Communication
Researchers have demonstrated a novel attack that subtly manipulates radio signals to compromise deep learning-based automatic modulation classifiers.
The Writing is on the Wall: Reimagining Education in the AI Era
As AI writing tools become increasingly sophisticated, educators must move beyond simply detecting plagiarism and focus on cultivating essential cognitive skills.
Truth Decay: How Well Do AI Models Spot Fake News?
A new study rigorously compares popular artificial intelligence techniques for detecting misinformation, revealing persistent challenges in generalizing across different news sources.
Spotting the Machine: A New Approach to AI Text Detection
Researchers have developed a novel, training-free method to reliably identify text generated by artificial intelligence by focusing on key linguistic signals within the text itself.
Simulating the Real World of High-Frequency Trading
![The quantile regression (QR) model failed to detect any discernible impact on the average price path surrounding a simulated metaorder, as evidenced by a distribution of inter-event times [latex]\Delta t[/latex] that aligns between empirical data (blue) and the QR prediction (green).](https://arxiv.org/html/2603.24137v1/x12.png)
New research tackles the challenges of accurately modeling limit order books to better evaluate trading strategies in dynamic markets.
Brain-Inspired AI Hones in on Deepfake Audio
Researchers are leveraging principles of neuroplasticity to create more efficient and accurate deepfake audio detection systems.
Splattered Reality: How Easily Can 3D Gaussian Splatting Be Fooled?
![Even imperceptible adversarial perturbations, generated through an attack like [latex]PGD[/latex], demonstrably degrade the reconstruction quality of a feed-forward 3D Gaussian Splatting model-specifically, the NoPoSplat implementation on the RE10K dataset-as evidenced by diminished performance metrics like normalized Peak Signal-to-Noise Ratio (PSNR) and qualitative visual artifacts in newly rendered views, despite utilizing multiple reference views.](https://arxiv.org/html/2603.23686v1/x1.png)
A new study reveals that even subtle, carefully crafted disturbances can significantly degrade the quality of reconstructions from feed-forward 3D Gaussian Splatting models.