Kevin Foley, Jonathan Hartadi, Shivesh Prakash, Charishma Takkallapalli, Swapnil Vatsal · 2026 · arXiv (Cornell University)
DOI: 10.48550/arxiv.2610.22745 · cited by 0 works (OpenAlex) · open-access PDF
Los modelos de idiomas que leen noticias financieras generalmente se entrenan en texto de los mismos años que se usan para probarlos, por lo que un backtest puede recompensar la vista atrás. Construimos temas de noticias de NoLBERT, un codificador pretrainado sólo en texto publicado entre 1976 y 1995, adaptado a un codificador de sentencia por contraste finamente ajustado en pasajes del mismo artículo, sin etiquetas ni retornos. Comparamos con los lectores de LDA, MiniLM y FinBERT en 1.26 millones de artículos de 2012 a 2024, convirtiendo los temas de cada lector en choques de atención y exposiciones a acciones, y prediciendo los retornos de las 1.000 acciones estadounidenses más grandes con ocho modelos en tres frecuencias de reequilibrio. Los modelos están equipados en 2013-2019, las opciones se hacen en
Machine translation (Meta M2M100); the English original follows.
Language models that read financial news are usually trained on text from the same years used to test them, so a backtest can reward hindsight. We build news topics from NoLBERT, an encoder pretrained only on text published between 1976 and 1995, adapted into a sentence encoder by contrastive fine-tuning on passages of the same article, with no labels or returns. We compare it with LDA, MiniLM and FinBERT readers on 1.26 million articles from 2012 to 2024, convert each reader's topics into attention shocks and stock exposures, and forecast the returns of the 1,000 largest US stocks with eight models at three rebalancing frequencies. Models are fitted on 2013-2019, choices are made on 2020-2021, and 2022-2024 is held out. The experiment with the best validation score loses money at every frequency. Measuring attention shocks against the past 63 trading days instead of 5 raises the test rank IC of every reader. NoLBERT gives the most stable topics and forecasts about as well as LDA. An equal-weight combination of all readers and models earns a monthly net Sharpe ratio of 1.12, significant against zero before adjusting for the search but not against a price-only benchmark.
Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese
Topics: Stock (firearms), Sentence, Econometrics, Computer science, Reading (process), Language model
GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.22745, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txtSources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange