February 2026

Spatio-temporal forecasting is fundamental to intelligent systems in transportation, climate science, and urban planning. However, training deep learning models on the massive, often redundant, datasets from these domains presents a significant computational bottleneck. Existing solutions typically focus on optimizing model architectures or optimizers, while overlooking the inherent inefficiency of the training data itself. This conventional approach of iterating over the entire static dataset each epoch wastes considerable resources on easy-to-learn or repetitive samples. In this paper, we explore […]

Ver mais

Class imbalance, overlap, and noise degrade data quality, reduce model reliability, and limit generalization. Although widely studied in binary classification, these issues remain underexplored in multi-class settings, where complex inter-class relationships make minority-majority structures unclear and traditional clustering fails to capture distribution shape. Approaches that rely only on geometric distances risk removing informative samples and generating low-quality synthetic data, while binarization approaches treat imbalance locally and ignore global inter-class dependencies. At the algorithmic level, ensembles struggle to integrate […]

Ver mais

Inverse lithography (ILT) is critical for modern semiconductor manufacturing but suffers from highly non-convex objectives that often trap optimization in poor local minima. Generative AI has been explored to warm-start ILT, yet most approaches train deterministic image-to-image translators to mimic sub-optimal datasets, providing limited guidance for escaping non-convex traps during refinement. We reformulate mask synthesis as conditional sampling: a generator learns a distribution over masks conditioned on the design and proposes multiple candidates. The generator is first pretrained […]

Ver mais

Detecting LLM training data is generally framed as a membership inference attack (MIA) problem. However, conventional MIAs operate passively on fixed model weights, using log-likelihoods or text generations. In this work, we introduce textbf{Active Data Reconstruction Attack} (ADRA), a family of MIA that actively induces a model to reconstruct a given text through training. We hypothesize that training data are textit{more reconstructible} than non-members, and the difference in their reconstructibility can be exploited for membership inference. Motivated by […]

Ver mais

I’ve written a couple times about Fibonacci numbers and certificates. Here the certificate is auxiliary data that makes it faster to confirm that the original calculation was correct. This post puts some timing numbers to this. I calculated the 10 millionth Fibonacci number using code from this post. n = 10_000_000 F = fib_mpmath(n) This took 150.3 seconds. As I’ve discussed before, a number f is a Fibonacci number if and only if 5f² ± 4 is a […]

Ver mais

I’m writing a series called “Roads to a Universal World Model”. I think this is arguably the most consequential open problem in AI and robotics right now, and most coverage either hypes it as “the next LLM” or buries it in survey papers. I’m trying to do something different: trace each major path from origin to frontier, then look at where they converge and where they disagree. The approach is narrative-driven. I trace the people and decisions behind […]

Ver mais