On the Position Bias of On-Policy Distillation
arXiv:2606.22600v3 Announce Type: replace-cross
Abstract: On-Policy Distillation (OPD) improves the learning efficiency of standard reinforcement learning through dense, token-level supervision from…
AI & Computational Science
The Remittance Blueprint: Data-driven Intelligence for Sri Lanka
arXiv:2606.28190v1 Announce Type: cross
Abstract: This study analyzes Sri Lankan migration and remittances over 32 years (1994-2025). Using a 384-month harmonized dataset, we apply exploratory data…
AI Predicts Time-Series Data Better by Analyzing Frequency Patterns
arXiv:2602.01588v3 Announce Type: replace-cross
Abstract: Multimodal time series forecasting is crucial in real-world applications, where decisions depend on both numerical data and contextual…
MultiHashFormer: Hash-based Generative Language Models
arXiv:2606.28057v1 Announce Type: cross
Abstract: Language models (LMs) represent tokens using embedding matrices that scale linearly with the vocabulary size. To constrain the parameter footprint,…
Complex-Valued 2D Gaussian Representation for Computer-Generated Holography
arXiv:2511.15022v2 Announce Type: replace-cross
Abstract: Complex-valued Gaussian primitives have recently been explored for representing holographic radiance fields in 3D novel view synthesis. In…
AI Learns to Sort Store Products Into Price Categories More Accurately
arXiv:2606.02004v2 Announce Type: replace-cross
Abstract: Consumer-price measurement increasingly draws on alternative data sources — scanner, web-scraped, and transaction/receipt data — whose…
What Is AI Security and Vulnerabilities? A Complete Guide to Understanding AI’s Achilles Heel
What Is AI Security and Vulnerabilities? A Complete Guide In 2022, researchers discovered something unsettling: they could fool state-of-the-art image recognition systems by adding imperceptible noise to photographs—noise so subtle that human eyes couldn’t detect it, yet it completely fooled the AI into misidentifying objects. A stop sign became a speed limit sign. A panda … Read more
Tuning Language Models by Mixture-of-Depths Ensemble
arXiv:2410.13077v2 Announce Type: replace-cross
Abstract: Transformer-based Large Language Models (LLMs) traditionally rely on final-layer loss for finetuning and final-layer representations for…
AI Framework Improves Time Series Predictions Using Context-Aware Dual Prototypes
arXiv:2601.16632v4 Announce Type: replace-cross
Abstract: Time series forecasting has witnessed significant progress with deep learning. While prevailing approaches enhance forecasting performance…
Efficient foundation decoders for fault-tolerant quantum computing
arXiv:2606.27119v1 Announce Type: cross
Abstract: Foundation decoders, a class of high-capacity neural decoders, are leading candidates for fault-tolerant quantum computing, with accurate and…