https://www.ml-quant.com/topics/llms-text/Large language models, agents, sentiment and text as data in finance. en https://www.ml-quant.com/papers/arxiv/2610.07798/https://www.ml-quant.com/papers/arxiv/2610.07798/Fri, 09 Oct 2026 07:00:00 +0000When financial facts are withdrawn from prompts, identity attributes explain 96% of variation in equity allocation advice from a language model, replacing missing evidence. https://www.ml-quant.com/papers/arxiv/2610.03080/https://www.ml-quant.com/papers/arxiv/2610.03080/Fri, 09 Oct 2026 07:00:00 +0000Introduces a benchmark that compares LLM-generated trading code to reference strategies bar by bar on identical data, revealing silent failures in implementation despite passing functional tests. https://www.ml-quant.com/papers/ssrn/7577906/https://www.ml-quant.com/papers/ssrn/7577906/Fri, 09 Oct 2026 07:00:00 +0000Shows that large-language-model scores of default risk from earnings calls predict bond rating migration one year ahead where analysts disagreed, capturing information not yet priced in spreads. https://www.ml-quant.com/papers/ssrn/7580558/https://www.ml-quant.com/papers/ssrn/7580558/Fri, 09 Oct 2026 07:00:00 +0000The study constructs a geoeconomic fear index from news articles and finds that higher fear predicts wider emerging-market sovereign spreads, especially for speculative-grade issuers. https://www.ml-quant.com/papers/arxiv/2609.35864/https://www.ml-quant.com/papers/arxiv/2609.35864/Fri, 02 Oct 2026 07:00:00 +0000Large language models outperform traditional methods at extracting missing financial data from SEC filings, with accuracy improving as model size matches document complexity. https://www.ml-quant.com/papers/repec/bri-uobdis-26-840/https://www.ml-quant.com/papers/repec/bri-uobdis-26-840/Fri, 02 Oct 2026 07:00:00 +0000Integrating financial news embeddings from large language models with market data, the study improves systemic risk forecasts using conditional value-at-risk without requiring large datasets. https://www.ml-quant.com/papers/repec/zbw-safewp-343061/https://www.ml-quant.com/papers/repec/zbw-safewp-343061/Fri, 02 Oct 2026 07:00:00 +0000Large language models provide nearly identical portfolio advice across twenty-one countries despite local differences, following retail finance conventions rather than academic prescriptions and ignoring household balance sheets. https://www.ml-quant.com/papers/arxiv/2609.23703/https://www.ml-quant.com/papers/arxiv/2609.23703/Fri, 25 Sep 2026 07:00:00 +0000Introduces a market-friction-aware framework that converts timestamped financial news into auditable trading decisions while accounting for execution timing, transaction costs, and liquidity constraints. https://www.ml-quant.com/papers/arxiv/2609.25034/https://www.ml-quant.com/papers/arxiv/2609.25034/Fri, 25 Sep 2026 07:00:00 +0000The study shows that how monetary policy sentiment unfolds across a press conference, not just its average tone, predicts rate changes and shapes forecaster expectations at the ECB and Fed. https://www.ml-quant.com/papers/arxiv/2609.24002/https://www.ml-quant.com/papers/arxiv/2609.24002/Fri, 25 Sep 2026 07:00:00 +0000A benchmark reveals that language models answer financial questions above 90 percent with clarification but only 28.9 percent when they must elicit it themselves, exposing model ambiguity resolution. https://www.ml-quant.com/papers/arxiv/2609.24175/https://www.ml-quant.com/papers/arxiv/2609.24175/Fri, 25 Sep 2026 07:00:00 +0000Develops a two-stage framework that fine-tunes language models for listwise asset ranking using Spearman rank correlation rewards, achieving a Sharpe ratio of 0.636 on asset allocation. https://www.ml-quant.com/papers/ssrn/7519200/https://www.ml-quant.com/papers/ssrn/7519200/Fri, 25 Sep 2026 07:00:00 +0000Semantic surprises extracted from Federal Reserve statements predict subsequent financial-stress dynamics and reduce forecast error by up to 23%, particularly when initial stress is high or during recessions. https://www.ml-quant.com/papers/arxiv/2503.01870/https://www.ml-quant.com/papers/arxiv/2503.01870/Thu, 16 Apr 2026 07:00:00 +0000Large Language Models are streamlining the process of identifying customer needs, letting analysts concentrate on more valuable work while still delivering precise insights. https://www.ml-quant.com/papers/ssrn/4467949/https://www.ml-quant.com/papers/ssrn/4467949/Sun, 28 Dec 2025 07:00:00 +0000A new financial sentiment index derived from Twitter data shows strong links to market conditions and can forecast stock market returns, particularly in response to changes in U.S. monetary policy. https://www.ml-quant.com/papers/arxiv/2512.09652/https://www.ml-quant.com/papers/arxiv/2512.09652/Sun, 14 Dec 2025 07:00:00 +0000An automated corruption index using Brazilian municipal audit reports is efficient and more reliable than manual methods in detecting corruption. https://www.ml-quant.com/papers/arxiv/2512.08270/https://www.ml-quant.com/papers/arxiv/2512.08270/Sun, 14 Dec 2025 07:00:00 +0000An evaluation of reasoning models on CFA mock exams shows that models like Gemini 3.0 Pro and GPT-5 perform well, achieving high pass rates in professional testing. https://www.ml-quant.com/papers/arxiv/2511.23057/https://www.ml-quant.com/papers/arxiv/2511.23057/Mon, 01 Dec 2025 07:00:00 +0000A project successfully developed a natural language processing model that classifies job ads with 72% accuracy by using an ensemble approach. https://www.ml-quant.com/papers/arxiv/2511.04299/https://www.ml-quant.com/papers/arxiv/2511.04299/Wed, 12 Nov 2025 07:00:00 +0000We build an interpretable, privacy‑friendly sentiment indicator from Swiss news using ML and LLMs, and it improves short‑term GDP forecasts. https://www.ml-quant.com/papers/arxiv/2510.19203/https://www.ml-quant.com/papers/arxiv/2510.19203/Mon, 27 Oct 2025 07:00:00 +0000Uses optimal-transport to align English–Japanese stock news, producing clearer signals that better predict returns. https://www.ml-quant.com/papers/arxiv/2510.20612/https://www.ml-quant.com/papers/arxiv/2510.20612/Mon, 27 Oct 2025 07:00:00 +0000LLM platforms can quietly absorb users’ ideas, creating power imbalances; the paper proposes governance and technical fixes to protect creators. https://www.ml-quant.com/papers/arxiv/2510.19799/https://www.ml-quant.com/papers/arxiv/2510.19799/Mon, 27 Oct 2025 07:00:00 +0000Combining transparent decision trees, LLMs, and practitioner input yields accurate, explainable case-level predictions for public and nonprofit programs. https://www.ml-quant.com/papers/arxiv/2510.19173/https://www.ml-quant.com/papers/arxiv/2510.19173/Mon, 27 Oct 2025 07:00:00 +0000- News-RL Trading - News-Driven Trading - News-Based RL - RL for News Trading - News-Powered RL If you want the shortest single choice: News-RL Trading.: Short summary: Feeding news sentiment extracted by large language models together with raw price and volume into a sequence-model reinforcement learning system boosts cryptocurrency trading performance, removing the need for handcrafted trading rules. https://www.ml-quant.com/papers/repec/spr-scient-v-130-y-2025-i-1-d-10-1007-s11192-024-05217-7/https://www.ml-quant.com/papers/repec/spr-scient-v-130-y-2025-i-1-d-10-1007-s11192-024-05217-7/Mon, 27 Oct 2025 07:00:00 +0000SVM vs BERT vs GPT-3.5: Compares SVM, SPECTER, BERT, and GPT-3.5 for classifying abstracts: BERT performs best, while GPT-3.5 is inconsistent with limited training data. https://www.ml-quant.com/papers/arxiv/2510.05475/https://www.ml-quant.com/papers/arxiv/2510.05475/Thu, 09 Oct 2025 07:00:00 +0000The potential of quantum logic in advancing artificial intelligence applications in financial modeling is discussed. https://www.ml-quant.com/papers/arxiv/2510.04357/https://www.ml-quant.com/papers/arxiv/2510.04357/Thu, 09 Oct 2025 07:00:00 +0000The research suggests the CSHT, a new architecture for financial time-series forecasting that considers the impact of financial news and sentiment on asset returns, providing robust generalisation across market regimes and clear attribution pathways. https://www.ml-quant.com/papers/doi/10-1145-3768292-3770344/https://www.ml-quant.com/papers/doi/10-1145-3768292-3770344/Fri, 03 Oct 2025 07:00:00 +0000The research explores the predictive power of textual features in earnings press releases on stock returns, concluding that press release content is as informative as earnings surprise, with FinBERT being the most predictive. https://www.ml-quant.com/papers/arxiv/2509.15510/https://www.ml-quant.com/papers/arxiv/2509.15510/Mon, 22 Sep 2025 07:00:00 +0000Large Language Models like ChatGPT have boosted earnings for workers in certain occupations without affecting unemployment rates, indicating they enhance income rather than replace jobs. https://www.ml-quant.com/papers/arxiv/2509.12519/https://www.ml-quant.com/papers/arxiv/2509.12519/Mon, 22 Sep 2025 07:00:00 +0000The study suggests using large language models to process financial news and small models to encode historical context, resulting in improved simulated investment performance. https://www.ml-quant.com/papers/arxiv/2509.11420/https://www.ml-quant.com/papers/arxiv/2509.11420/Mon, 22 Sep 2025 07:00:00 +0000The article presents Trading-R1, a finance-focused AI model that aligns with trading principles, showing it offers better risk-adjusted returns and fewer drawdowns than other models. https://www.ml-quant.com/papers/arxiv/2508.17906/https://www.ml-quant.com/papers/arxiv/2508.17906/Fri, 29 Aug 2025 07:00:00 +0000Financial Knowledge Graph Construction: The paper introduces a large-scale financial knowledge graph dataset from SEC 10-K filings of S and P 100 companies, with a reflection-agent-based mode providing the best balance of efficiency, accuracy, and reliability.