Skip to content

Research separate from news

What could become important next

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.

1381 published research events

Latest work

A significant claim from a single source is published only after further confirmation.

Research only one source so far

New study proposes a statistically grounded method for choosing a noise schedule for training diffusion models

A study on arXiv presents a statistical framework for the optimal distribution of noise levels when training diffusion models. According to the authors, an entropy-derived schedule significantly improves training efficiency in discrete domains and is competitive with standard EDM heuristics for images.

arXiv cs.LG (Machine Learning) Original source ↗
Research only one source so far

FlowEdit: Managing conflicting LLM responses using information theory

A research paper proposes FlowEdit, a method for addressing ill-posed problems (inconsistent conditions with no solution) by controlling internal LLM reasoning flows. The method generates alternative responses for competing hypotheses. Experiments show a 68 % improvement in exact-set-match accuracy and a 24 % improvement in…

arXiv cs.AI (Artificial Intelligence) Original source ↗
Research only one source so far

ExecuGraph: A verification framework for generating backend code with LLMs

A study introduces ExecuGraph, a multi-agent framework for backend code synthesis with LLMs. Six agents (Planner, Code Generator, Logical Reviewer, Evaluator, Optimizer, Explainer) collaborate in a typed workflow with isolated sandbox evaluation. On DeepSeekCoder V2 Lite, it achieved an accuracy increase of 22.5…

arXiv cs.AI (Artificial Intelligence) Original source ↗
Research only one source so far

Benchmarking personalization capabilities in large language models

A research study adapts the Bayesian Persuasion framework for generative agents and tests LLM personalization capabilities in sales. The authors released SDR-Bench, a dataset with 6 279 cases from 22 industries, and found a consistent personalization plateau — among Fortune 100 tech companies, no model differs statistically in…

arXiv cs.AI (Artificial Intelligence) Original source ↗
Research only one source so far

LeanFlow: Formalizing mathematical texts using LLM agents

LeanFlow automates the translation of mathematical publications into the Lean formal language. Tested on papers in number theory and measure theory; Kimi2.6 completes both projects within a budget of 2000 API calls, while GPT5.5 achieves the lowest token costs. It solves all five tasks from ICML 2026 AI for Math TCS…

arXiv cs.AI (Artificial Intelligence) Original source ↗
Research only one source so far

Training-free routing: Managing LLM offloading through a reliability gate

A research team proposes CARGO, a training-free framework for dynamic routing between a local and a cloud LLM. The method uses prompt-varied sampling and Bayesian early stopping to estimate the local model's reliability without additional router training. In tests, it outperformed some supervised…

arXiv cs.AI (Artificial Intelligence) Original source ↗
Research only one source so far

From errors to rules: Iterative prompt optimization for text classification

A research study compares prompt optimization approaches (demonstration selection, exploration, error diagnosis) for text classification. It proposes ERGO, a method that iteratively diagnoses errors and generates decision rules. It achieves the best results on tasks with learnable boundaries: TREC 90…

arXiv cs.AI (Artificial Intelligence) Original source ↗
Research only one source so far

CT-Merging: A new algorithm for merging LoRA adapters

A research preprint describes CT-Merging, an algorithm for efficiently merging multiple LoRA adapters into a single universal adapter. On the DC-Merge CLIP benchmark, it improves by 2.56 points on ViT-B/32 and 1.51 points on ViT-L/14 over existing methods.

arXiv cs.LG (Machine Learning) Original source ↗
Research only one source so far

VeriSimpl: Natural language optimization with simplification-based verification

A research paper describes VeriSimpl, a framework for translating natural language into optimization models using LLMs. Its innovation is "simplification-based verification" — an optimization solver generates simplified queries to verify that the formulation is correct. The method achieves higher accuracy than existing…

arXiv cs.AI (Artificial Intelligence) Original source ↗
Research only one source so far

Aligning language model expectations with real user needs

A research study examines how well large language models align with what users actually want. The authors introduce ExpectBench, a benchmark with real user expectations, and the LENS framework for generating better-aligned responses.

arXiv cs.AI (Artificial Intelligence) Original source ↗
Research only one source so far

SonicSampler: Unified kernels for sampling and speculative verification in LLMs

A research paper introduces optimized Triton kernels for accelerating LLM inference. SonicSampler combines logit processing and token selection into a single kernel, supporting dynamic behavior per request. It achieves speedups of up to 10× for the top-k algorithm and up to 16× for speculative decoding while preserving…

arXiv cs.AI (Artificial Intelligence) Original source ↗
Research only one source so far

JAXBench: A benchmark for autonomous TPU kernel optimization

A study introduces JAXBench, a benchmark suite with 50 JAX workloads for optimizing TPU kernels using AI. It includes 17 production operators from models such as Llama-3.1 and DeepSeek-V3. With Gemini 3 Flash, correctness increased from 5.8% to 37.3%, with speedups of 1.28x to 1.36x.

arXiv cs.AI (Artificial Intelligence) Original source ↗
Research only one source so far

EvoSQL: Co-evolution of a generator and a critic with memory for text-to-SQL

A research paper introduces EvoSQL, a method for improving SQL generation from natural language. It combines a generator and an LLM critic with candidate memory, verifies SQL through execution and LLM analysis, and improves through SDPO fine-tuning. Tests on the Spider and BIRD benchmarks show improvements; on BIRD-Dev, performance…

arXiv cs.AI (Artificial Intelligence) Original source ↗
Research only one source so far

Deep learning-based surrogate models for predicting electrode behavior in lithium-ion batteries

Scientists developed a deep learning pipeline with a Swin3D Transformer to predict the spatiotemporal behavior of electrodes in Li-ion batteries. Gaussian Positional Encoding and Temporal Encoding innovations improve accuracy over point-cloud methods. The method reduces computational demands by orders of magnitude and enables faster…

arXiv cs.LG (Machine Learning) Original source ↗
Research only one source so far

Explanation-based runtime verification for reliable optical networks using ML

A scientific preprint describes an approach to runtime verification of machine learning model decisions in optical networks. The system checks the coherence of model explanations and the physical correctness of decisions before executing them in automated control loops. Experiments on optical link quality classification…

arXiv cs.LG (Machine Learning) Original source ↗