Research separate from news
A preprint is a signal, not a finished product or an independently confirmed result. We therefore track research papers separately and show the actual state of supporting evidence for each one.
A significant claim from a single source is published only after further confirmation.
A research paper describes SiGMA, a method that addresses negative interference and forgetting in continual learning with multimodal LLMs. It combines sign-guided adaptive tuning and sign-guided merging. Tests on the UCIT and DCL benchmarks show improvements over previous methods.
A research study develops a Deep Q-Network framework for autonomous conflict resolution between heterogeneous unmanned aircraft and EVTOL in 3D corridors. Tests across 90 traffic density combinations showed that conflicts are usually resolved within 1 second and agents maintain their heading 79 % of the time.
An arXiv position paper argues that reactively fixing AI models based on user-reported errors has limitations and should be replaced by a proactive test-driven flywheel. The authors mathematically prove that the proactive approach achieves better long-term scaling with fewer iterations.
A research paper introduces SalesLoop, a method for ranking potential customers in CRM systems using reinforcement learning based on actual sales outcomes. Tests show a +7.9 % increase in NDCG@K and +15.8 % in P@K, with a production A/B test at a Chinese electric vehicle manufacturer with 16.5 million leads…
A research paper introduces DecodeShare, a protocol for identifying a low-dimensional subspace shared across tasks in hidden states during LLM decoding. Disrupting this subspace degrades performance more than disrupting a prefill-derived or random subspace, with implications for activation…
A research team developed a decision-aware machine learning system for allocating essential medicines in low-income countries. A pilot deployment in Sierra Leone with the national government showed a 19% increase in consumption of allocated products and reached 2 million women and children under five.
A study proposes weighting multiple language models by the quality of their uncertainty rather than assuming equal trustworthiness. A Cooke-style method penalizes confidently wrong predictions. Tests on MMLU and MMLU-Pro showed robustness to unreliable models in heterogeneous panels.
Research showed that Monte Carlo dropout applied to an X-ray classifier improves detection of incorrect diagnoses (AUROC +0.023). When the uncertainty signal was presented as a binary warning rather than numerical scores, doctors reduced confidently wrong diagnoses from 8.5 % to 2.7 %. Uncertainty carries decision value…
The PlanE research framework optimizes datasets for tailoring LLMs to specific tasks with the aim of reducing annotation costs. It includes a DTI planning algorithm for selecting the optimal model and parameter combinations. The code is publicly available.
A mathematical study of how neural networks with the ReLU activation function can be constructed to exactly implement certain iterative mathematical operations. The main contribution is a new residual memory controller algorithm, which enables efficient implementation with a depth of O(n).
Researchers propose a reinforcement learning method that adaptively selects and combines multiple time horizons instead of a fixed discount factor. It enables better adaptation to changes in the reward structure without manual tuning. The method was tested in MiniGrid environments, including continual learning with…
A research paper on arXiv introduces the PersonaTrail benchmark for evaluating web agents capable of working with underspecified instructions and inferring context from browsing history. It includes the Preference-Aware Contextual Memory (PACMem) framework, which distinguishes between factual and preference memories for…
A study tested 5 watermarking schemes (tracing marks) on 11 LLMs and 7 VLMs in medical tasks. It found that watermarks cause performance degradation — lexical errors, hallucinated specialist terminology and errors in interpreting medical images. Aggregate metrics usually hide these problems.
A research paper describes a method for training neural networks to control complex systems. It combines control barrier functions with learning-based feedback control, enabling scalable training on problems with up to 1200 state dimensions and up to 400 control dimensions.
A research paper on arXiv introduces SevDiff, a diffusion model for generating realistic vehicle trajectories with precisely controlled collision severity (Time-to-Collision). The model achieves 100% accuracy for TTC 0.5–1.5 s and 97–99% for 2.0–2.5 s, trained on 468 interaction sequences from express…
A scientific article on CEDAR, a method for detecting causal edges in sparse autoregressive time series. It combines screening of candidate lags using U-centered distance correlation with conditional independence tests, requires O(d²) tests and is most effective in data-scarce situations.
The research examines how transformers with weighted layers applied T times implement algorithms. The authors identify four key findings: (1) budget law — a linear computational bound related to training conditions (v ~ n_train/T_train), (2) architecture determines the type of algorithm (parallel vs.…
An academic preprint proposes a formal approach to the Black-Litterman portfolio construction model that uses neural predicates to automatically generate investor views and their uncertainty. The model is interpretable and fully differentiable, enabling end-to-end learning from financial data.
AI Radar monitors Czech and international sources every day, looking for changes that truly deserve attention.
The result is a quick overview of what has actually changed in the AI world, rather than another stream of articles.
Use the CS/EN switch to read the same Radar in Czech or English. English content is published after its translation has been checked, so new and older items may appear later.
Each row represents one event — not one article. At a glance, you can see its significance, credibility and main point.
Illustrative example, not a current news item.
Agent mode is available to all paying users
Until now, the mode was available only on the highest plan; it is now available on all paid tiers without a waitlist.
The detail page contains a fuller summary, its significance and original sources. Practical impact appears in the detail and the For individuals and For businesses views. An AI Flash item reaches the main selection only after it has been expanded and meets the publication rules.
We first summarize each event in the same way for everyone. Based on those same facts, we then explain what the change means for your own use and what it could mean for how a company operates.
Today’s briefing is the same for everyone. Pages For individuals and For businesses can be found in the main navigation — they select only events relevant to the given use case.
Reset history · Outage history · Subscribe in a feed reader (RSS) · Methodology and label explanations
