Tootfinder

LLMs are now part of our daily work, making coding easier. Join Ivan Dolgov at this year's Berlin Buzzwords to learn how they built an in-house LLM for AI code completion in JetBrains products, covering design choices, data preparation, training and model evaluation.
Learn more: https://

Chance and Mass Interpretations of Probabilities in Markov Decision Processes (Extended Version)
Yun Chen Tsai, Kittiphon Phalakarn, S. Akshay, Ichiro Hasuo
https://arxiv.org/abs/2506.10377

Chance and Mass Interpretations of Probabilities in Markov Decision Processes (Extended Version)
Markov decision processes (MDPs) are a popular model for decision-making in the presence of uncertainty. The conventional view of MDPs in verification treats them as state transformers with probabilities defined over sequences of states and with schedulers making random choices. An alternative view, especially well-suited for modeling dynamical systems, defines MDPs as distribution transformers with schedulers distributing probability masses. Our main contribution is a unified semantical framew…

EXPO: Stable Reinforcement Learning with Expressive Policies
Perry Dong, Qiyang Li, Dorsa Sadigh, Chelsea Finn
https://arxiv.org/abs/2507.07986 https://arxiv.org/pdf/2507.07986 https://arxiv.org/html/2507.07986
arXiv:2507.07986v1 Announce Type: new
Abstract: We study the problem of training and fine-tuning expressive policies with online reinforcement learning (RL) given an offline dataset. Training expressive policy classes with online RL present a unique challenge of stable value maximization. Unlike simpler Gaussian policies commonly used in online RL, expressive policies like diffusion and flow-matching policies are parameterized by a long denoising chain, which hinders stable gradient propagation from actions to policy parameters when optimizing against some value function. Our key insight is that we can address stable value maximization by avoiding direct optimization over value with the expressive policy and instead construct an on-the-fly RL policy to maximize Q-value. We propose Expressive Policy Optimization (EXPO), a sample-efficient online RL algorithm that utilizes an on-the-fly policy to maximize value with two parameterized policies -- a larger expressive base policy trained with a stable imitation learning objective and a light-weight Gaussian edit policy that edits the actions sampled from the base policy toward a higher value distribution. The on-the-fly policy optimizes the actions from the base policy with the learned edit policy and chooses the value maximizing action from the base and edited actions for both sampling and temporal-difference (TD) backup. Our approach yields up to 2-3x improvement in sample efficiency on average over prior methods both in the setting of fine-tuning a pretrained policy given offline data and in leveraging offline data to train online.
toXiv_bot_toot

Revealing Dark Matter's Role in Neutron Stars Anisotropy: A Bayesian Approach Using Multi-messenger Observations
Xue-Zhi Liu, Premachand Mahapatra, Chun Huang, Ayush Hazarika, Chiranjeeb Singha, Prasanta Kumar Das
https://arxiv.org/abs/2506.08376

Revealing Dark Matter's Role in Neutron Stars Anisotropy: A Bayesian Approach Using Multi-messenger Observations
Dark matter (DM) continues to evade direct detection, but neutron stars (NSs) serve as natural laboratories where even a modest DM component can alter their structure. While many studies have examined DM effects on NSs, they often rely on specific choices of equations of state (EOS) models, assume isotropy, and lack a Bayesian statistical framework, limiting their predictive power. In this work, we present a Bayesian framework that couples pressure-anisotropic nuclear EOS to a self-interacting …

When Trump encounters a demographic he doesn’t care for, he “disappears” them–in many cases, the very record of their existence.
Meet America’s new “missing persons," and the govt campaign to delete statistical evidence of anyone who doesn't fit Trump’s idea of our new golden age... including Jackie Robinson & Rosa Parks.
(taken from Catherine Rampell on BlueSky)
☑️ The people Trump is trying to delete from government records - Washington Post

Weighted Parameter Estimators of the Generalized Extreme Value Distribution in the Presence of Missing Observations
James H. McVittie, Orla A. Murphy
https://arxiv.org/abs/2506.15964

Weighted Parameter Estimators of the Generalized Extreme Value Distribution in the Presence of Missing Observations
Missing data occur in a variety of applications of extreme value analysis. In the block maxima approach to an extreme value analysis, missingness is often handled by either ignoring missing observations or dropping a block of observations from the analysis. However, in some cases, missingness may occur due to equipment failure during an extreme event, which can lead to bias in estimation. In this work, we propose weighted maximum likelihood and weighted moment-based estimators for the generaliz…

A Data Science Approach to Calcutta High Court Judgments: An Efficient LLM and RAG-powered Framework for Summarization and Similar Cases Retrieval
Puspendu Banerjee, Aritra Mazumdar, Wazib Ansar, Saptarsi Goswami, Amlan Chakrabarti
https://arxiv.org/abs/2507.01058

A Data Science Approach to Calcutta High Court Judgments: An Efficient LLM and RAG-powered Framework for Summarization and Similar Cases Retrieval
The judiciary, as one of democracy's three pillars, is dealing with a rising amount of legal issues, needing careful use of judicial resources. This research presents a complex framework that leverages Data Science methodologies, notably Large Language Models (LLM) and Retrieval-Augmented Generation (RAG) techniques, to improve the efficiency of analyzing Calcutta High Court verdicts. Our framework focuses on two key aspects: first, the creation of a robust summarization mechanism that distills…

Reinforcement Learning for Optimal Control of Spin Magnetometers
Logan W. Cooke, Stefanie Czischek
https://arxiv.org/abs/2506.21475 https://

Reinforcement Learning for Optimal Control of Spin Magnetometers
Quantum optimal control in the presence of decoherence is difficult, particularly when not all Hamiltonian parameters are known precisely, as in quantum sensing applications. In this context, maximizing the sensitivity of the system is the objective, for which the optimal target state or unitary transformations are unknown, especially in the case of multi-parameter estimation. Here, we investigate the use of reinforcement learning (RL), specifically the soft actor-critic (SAC) algorithm, for pr…

excellent analysis of covid state of play
From Julia Doubleday.
"Each time a new wave crops up, the media scrambles to let the public know that COVID is spreading “again” ... But each time, it fails to inform the public that nearly half of COVID cases are asymptomatic, that COVID looks different in different patients, that vaccines do not prevent infections, that rapid tests have high false negative rates, and that COVID is fully airborne.
"Taken together, the virus I’m describing is much more difficult to control than the one the press presents. The press frames the virus as something that can be halted by familiarizing yourself with the symptoms, staying home once you feel sick and test positive, and avoided altogether by simply getting vaccinated or keeping ones’ distance from sick people. ...
"The misinformation that reigns in liberal spaces is not the result of accidental miscommunication. People don’t know that the virus is asymptomatic 40% of the time because there is simply no universe where that virus is controllable without an elimination strategy, or a day-to-day mitigation strategy."
#covid #misinformation #denial #CovidIsAirborne

Constructing Evidence-Based Tailoring Variables for Adaptive Interventions
John J. Dziak, Inbal Nahum-Shani
https://arxiv.org/abs/2506.03054 https://

Constructing Evidence-Based Tailoring Variables for Adaptive Interventions
Background: An adaptive intervention (ADI) uses individual information in order to select treatment, to improve effectiveness while reducing cost and burden. ADIs require tailoring variables: person- and potentially time-specific information used to decide whether and how to deliver treatment. Specifying a tailoring variable for an intervention requires specifying what to measure, when to measure it, when to make the resulting decisions, and what cutoffs should be used in making those decisions…