01 DAILY RESEARCH SIGNALS / 2026-10-11
A curated stream of quantitative research, open-source projects, and AI engineering advances—screened, ranked, and structured for faster research decisions.
All sources checked. Pipeline operating normally.
02 CURATED INDEX
Browse by content type or search titles and tags.
INDEXED 15
SELECTED SIGNALS 03
The project connects real open-source spatial intelligence data to a browser-based 3D globe, offering a satellite-reconnaissance-like visualization; it is novel in combining live data with a photorealistic 3D globe as an interactive tool, useful for developers interested in geospatial visualization.
The GitHub trending project rea proposes using agents for reverse engineering, covering everything from app behavior down to native binaries. By delegating reverse-engineering workflows to agents, it may offer a more automated path for security analysis and program understanding.
Open WebUI is a user-friendly AI interface that supports backends such as Ollama and the OpenAI API. It is new in offering a unified interaction entry compatible with multiple model services, simplifying daily use of local or cloud-based large models; it is practical for users who need a self-hosted AI chat entry.
SELECTED SIGNALS 11
This work proposes structured disambiguation, where users resolve ambiguity through explicit, constrained interactions instead of free-form dialogue clarification, and introduces the ARCS benchmark, which annotates naturally occurring ambiguity points, interpretations, and SQL queries over real-world databases. Its value lies in turning ambiguity from a conversational burden into a reproducible input condition for evaluating intent misalignment in text-to-SQL deployment.
Existing methods typically fix the optimizer's own tools and procedures, improving only the target agent's prompts, tools, and workflow. VERSE enables the optimizer to also improve how it diagnoses failures, develops edits, and tests effects, using execution-based verification, failure replay, and step perturbation to enhance optimization. On SWE-rebench in-distribution and cross-language out-of-distribution tasks, its best validation-selected harness reaches 42.3% and 37.7%
This work introduces a test-time framework that, without retraining policies, combines multiple frozen goal-conditioned policies into a portfolio, selecting the best policy at each state by imagining futures in a world model. Its novelty lies in comparing heterogeneous policies via a shared value function and avoiding excessive switching. On 18 state-based OGBench datasets, it improves macro-average success from 44% to 58%, offering a plug-and-play policy integration approach
Existing agent safety evaluations mostly focus on digital environments. This work introduces RT-SAFE, a simulated urban benchmark in which the world keeps evolving during inference and action execution, bringing decision latency into embodied safety assessment. Its value lies in showing that high task completion can mask safety failures and in offering an environment for offline RL training to reduce collisions.
This study addresses first-pass failures in locally deployed LLM-generated power-system simulation code, proposing a deployment-time workflow that requires no weight updates: documentation-driven L0-L3 probes build per-model API profiles, query-side demand estimation selects layered API evidence before generation, and execution feedback routes targeted repair. Its value lies in turning common API hallucination and parameter misuse into a diagnosable and calibratable engineeri
Formalizes the systemic risk of agentic AI across six mathematical settings: from an expected-harm identity for expected loss and the non-diversifiable common exposure of a shared foundation model to an observability trichotomy and runtime-impossibility corollary for runtime guardrails, arguing the systematic component of agentic risk cannot be diversified away, detected away, or reversed and can only be met by ex-ante structural prevention.
These lecture notes introduce optimal transport as a mathematical language for actuarial science, treating losses, premiums, reserves, and climate losses as probability measures to be compared, transported, averaged, stressed, and interpolated. The novelty lies in systematically connecting optimal transport tools with actuarial problems such as risk measures, Wasserstein robustness, pricing, capital allocation, and longevity risk, while emphasizing the interpretation of model
DSTNet introduces a causal spectral trajectory representation that retains the recent evolution of wavelet filter-bank magnitudes rather than only denoising or using a single-point snapshot, combined with a scale-temporal factored Transformer and gated fusion. On seven equity indices and gold under an expanding-window protocol, the model beats the random-walk persistence benchmark in all 32 series-horizon cells, but paired testing is inconclusive against the strongest competi
The work introduces the two-particle-irreducible effective action and Dyson-Schwinger gap equations from quantum field theory into stochastic-volatility option pricing as a non-perturbative closure. Its novelty lies in replacing Taylor truncation with the exact Gaussian moment-generating function for exponential and CEV vertices; in exp-OU the melon-level correction reduces the Hartree error by 20-130x to 0.01-0.11 bp. The method targets calibration, forward-start smiles, bar
This study proposes a stochastic portfolio theory framework in which trading strategies are generated from nominal price weights and evaluated against a price-weighted benchmark. Its novelty lies in incorporating split- and reverse-split-induced jumps in generating weights into the self-financing condition and providing explicit jump-corrected wealth decompositions; it is useful because it shows that the usual long-horizon relative arbitrage argument does not extend directly
The study proposes a Dirichlet mixed-membership model that extends classical credibility theory from conditional means to multivariate distributions, enabling insurers to combine portfolio priors and individual experience posteriors exactly.
SELECTED SIGNALS 01
This card is based on Quantocracy's link digest dated Oct 10, 2026, which includes a discussion on the role of long bonds in portfolios; however, the given content only mentions that the 30-year Treasury yield closed at 5.60% in the first week of October, its highest since 2002, and then truncates after 'TLT', without providing specific new findings or conclusions.