Federated Learning (FL) enables distributed learning across multiple clients without sharing raw data. When statistical heterogeneity across clients is severe, Clustered Federated Learning (CFL) can improve performance by grouping similar clients and training cluster-wise models. However, most CFL approaches rely on multiple communication rounds for cluster estimation and model updates, which limits their practicality under tight constraints on communication rounds. We propose Data Collaboration-based Clustered Federated Learning (DC-CFL), a single-round framework that completes both client clustering and cluster-wise […]

Ver mais

Your Computer System Just Became Your Universal AI Agent Toolkit For Everything! :::tip What if you could talk to your computer the way you talk to a colleague — and it just did the work? ::: Not autocomplete. Not a chatbot that pastes code snippets. Not a suggestion engine that highlights the next three words. A system that takes your intent and orchestrates your entire machine — files, browser, terminal, internet — to deliver the result. That is exactly […]

Ver mais

I’ve been thinking about an approach for fine-tuning/RL on limited data and I’m not sure it’s the right one , curious if anyone has done something similar. i need a model that generates document templates from structured input + a nl comment. The only data I have are existing compiled templates, no input/output pairs. The idea is to bootstrap with reverse engineering, feed each template to a strong LLM, extract the parameters that could have generated it, use […]

Ver mais

arXiv:2504.15543v3 Announce Type: replace-cross Abstract: Computer Adaptive Testing (CAT) aims to accurately estimate an individual’s ability using only a subset of an Item Response Theory (IRT) instrument. Many applications also require diverse item exposure across testing sessions, preventing any single item from being over- or underutilized. In CAT, items are selected sequentially based on a running estimate of a respondent’s ability. Prior methods almost universally see item selection through an optimization lens, motivating greedy item selection procedures. While […]

Ver mais

arXiv:2603.06431v1 Announce Type: cross Abstract: Neural network methods for PDEs require reliable error control in function space norms. However, trained neural networks can typically only be probed at a finite number of point values. Without strong assumptions, point evaluations alone do not provide enough information to derive tight deterministic and guaranteed bounds on function space norms. In this work, we move beyond a purely black-box setting and exploit the neural network structure directly. We present a framework for […]

Ver mais

arXiv:2602.13310v1 Announce Type: new Abstract: Existing LLM test-time scaling laws emphasize the emergence of self-reflective behaviors through extended reasoning length. Nevertheless, this vertical scaling strategy often encounters plateaus in exploration as the model becomes locked into specific thinking pattern. By shifting from depth to parallelism, parallel thinking mitigates the narrowing of exploration. However, the extension of this paradigm to visual domain remains an open research question. In this paper, we first examine the role of visual partitioning in […]

Ver mais

arXiv:2603.06697v1 Announce Type: new Abstract: Vision–language models (VLMs) process images as visual tokens, yet their intermediate reasoning is often carried out in text, which can be suboptimal for visually grounded radiology tasks. Radiologists instead diagnose via sequential visual search; eye-tracking captures this process as time-ordered gaze trajectories that reveal how evidence is acquired over time. We use eye-gaze as supervision to guide VLM reasoning by introducing a small set of dedicated gaze tokens. These tokens are trained to […]

Ver mais

The Clawdbot platform has quickly become the most popular AI chatbot system that businesses use to deliver customer support, execute internal operations, and manage advanced business processes. The bots now serve as critical data protection systems as organizations adopt AI technologies, which require these systems to handle confidential information. Existing protection systems cannot manage AI chatbot security threats because they lack the capabilities to address prompt manipulation and API security breaches. Organizations must comprehend these security dangers because […]

Ver mais

Hi everyone, I’ve written a short research note on GitHub where I explore an idea related to making LLMs improve their own training process by self-distribution aware analysis. The focus is not on a specific implementation, but on a general training paradigm and how models could guide what data or signals they learn from next. I’m looking for feedback or criticism. My goal is discussion and learning, not making any strong claims. If someone finds the direction interesting […]

Ver mais

arXiv:2601.15503v1 Announce Type: new Abstract: Volunteer-led lake monitoring yields irregular, seasonal time series with many gaps arising from ice cover, weather-related access constraints, and occasional human errors, complicating forecasting and early warning of harmful algal blooms. We study Secchi Disk Depth (SDD) forecasting on a 30-lake, data-rich subset drawn from three decades of in situ records collected across Maine lakes. Missingness is handled via Multiple Imputation by Chained Equations (MICE), and we evaluate performance with a normalized Mean […]

Ver mais