Searcharxiv⌕ Search

arXiv subjects

Kevin Foley

Publications and source records attributed to Kevin Foley.

7 recordsLinked to original sources

NoLBERT-Calls: Reading Earnings Calls Without Hindsight

Language models are now a standard tool for measuring what managers say on earnings calls, but the encoders behind these measures were usually pretrained on text from the same years a backtest covers. Holding out a test period does not remove that hindsight, because it sits in the features. We adapt NoLBERT, an encoder pretrained only on text published between 1976 and 1995, to earnings calls without letting it read anything dated after 31 December 2023. The calls supply the training signal themselves: every analyst question is followed by management's answer, and the model learns to pick that answer out from the other answers of the same call. We test on every US call from January 2025 to October 2026 (20,323 calls, 254,575 questions) and make every modeling choice on calls from 2024. The adapted model, NoLBERT-Calls-2023, ranks the true answer first for 78.0% of questions (95% interval 77.8-78.2; chance 8.0%), against 26.3% before adaptation. It is 1.0 point ahead of a call-adapted ChronoBERT, 3.4 points ahead of a call-adapted FinBERT, and 18.7 and 19.8 points ahead of the best off-the-shelf retriever and BM25. A year-by-year test over 2013-2026 finds no look-ahead premium: relative to a model adapted only through 2012, the released model does no better on the years it was trained on than on years it never saw. The price of the cutoff is compute: under one shared 150-minute recipe NoLBERT adapts least well of four encoders, and its lead comes from a longer, staged adaptation of about 13 GPU-hours.

cs.CL↗

Reading the News Without Hindsight: NoLBERT Topics and the Cross-Section of Stock Returns

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.

q-fin.ST↗

From Word Counts to Context: Topic Models for Asset Pricing

News may reveal systematic risk, but whether its context enhances the construction of systematic risk factors is still unclear. We seek to test whether utilizing a sentence transformer represents an improvement over techniques such as Latent Dirichlet Allocation (LDA) in the coherence of topic term lists generated from unstructured text data. To test this, the same collection of unstructured text data comprising of 394,661 articles and the same downstream financial portfolio construction pipeline were applied with the text layer differing, including the length of article text each model used and how topic terms were ranked: we benchmark LDA against a frozen sentence transformer with k-means clustering. We find that the sentence transformer branch had higher observed scores both in terms of coherence (measured by NPMI) as well as financial performance (measured by Sharpe), although the available tests do not establish outperformance. Further exploratory specifications such as utilizing spherical clustering and multi-horizon exposures had an observed excess-return Sharpe of 1.03 for the combined model. We believe that there is some promise in applying context-aware techniques on unstructured news text, but stricter tests using only information available at each date and broader datasets may be required to enhance the confidence in the observed performance.

cs.CE↗

Architecture is All You Need: Improving LLM Recommenders by Dropping the Text

In recent years, there has been an explosion of interest in the applications of large pre-trained language models (PLMs) to recommender systems, with many studies showing strong performance of PLMs on common benchmark datasets. PLM-based recommender models benefit from flexible and customizable prompting, an unlimited vocabulary of recommendable items, and general ``world knowledge'' acquired through pre-training on massive text corpora. While PLM-based recommenders show promise in settings where data is limited, they are hard to implement in practice due to their large size and computational cost. Additionally, fine-tuning PLMs to improve performance on collaborative signals may degrade the model's capacity for world knowledge and generalizability. We propose a recommender model that uses the architecture of large language models (LLMs) while reducing layer count and dimensions and replacing the text-based subword tokenization of a typical LLM with discrete tokens that uniquely represent individual content items. We find that this simplified approach substantially outperforms both traditional sequential recommender models and PLM-based recommender models at a tiny fraction of the size and computational complexity of PLM-based models. Our results suggest that the principal benefit of LLMs in recommender systems is their architecture, rather than the world knowledge acquired during extensive pre-training.

cs.IR↗

Predicting Movie Hits Before They Happen with LLMs

Addressing the cold-start issue in content recommendation remains a critical ongoing challenge. In this work, we focus on tackling the cold-start problem for movies on a large entertainment platform. Our primary goal is to forecast the popularity of cold-start movies using Large Language Models (LLMs) leveraging movie metadata. This method could be integrated into retrieval systems within the personalization pipeline or could be adopted as a tool for editorial teams to ensure fair promotion of potentially overlooked movies that may be missed by traditional or algorithmic solutions. Our study validates the effectiveness of this approach compared to established baselines and those we developed.

cs.IR↗

Improving Content Recommendation: Knowledge Graph-Based Semantic Contrastive Learning for Diversity and Cold-Start Users

Addressing the challenges related to data sparsity, cold-start problems, and diversity in recommendation systems is both crucial and demanding. Many current solutions leverage knowledge graphs to tackle these issues by combining both item-based and user-item collaborative signals. A common trend in these approaches focuses on improving ranking performance at the cost of escalating model complexity, reducing diversity, and complicating the task. It is essential to provide recommendations that are both personalized and diverse, rather than solely relying on achieving high rank-based performance, such as Click-through Rate, Recall, etc. In this paper, we propose a hybrid multi-task learning approach, training on user-item and item-item interactions. We apply item-based contrastive learning on descriptive text, sampling positive and negative pairs based on item metadata. Our approach allows the model to better understand the relationships between entities within the knowledge graph by utilizing semantic information from text. It leads to more accurate, relevant, and diverse user recommendations and a benefit that extends even to cold-start users who have few interactions with items. We perform extensive experiments on two widely used datasets to validate the effectiveness of our approach. Our findings demonstrate that jointly training user-item interactions and item-based signals using synopsis text is highly effective. Furthermore, our results provide evidence that item-based contrastive learning enhances the quality of entity embeddings, as indicated by metrics such as uniformity and alignment.

cs.IR↗

Blending search queries with social media data to improve forecasts of economic indicators

The forecasting of political, economic, and public health indicators using internet activity has demonstrated mixed results. For example, while some measures of explicitly surveyed public opinion correlate well with social media proxies, the opportunity for profitable investment strategies to be driven solely by sentiment extracted from social media appears to have expired. Nevertheless, the internet's space of potentially predictive input signals is combinatorially vast and will continue to invite careful exploration. Here, we combine unemployment related search data from Google Trends with economic language on Twitter to attempt to nowcast and forecast: 1. State and national unemployment claims for the US, and 2. Consumer confidence in G7 countries. Building off of a recently developed search-query-based model, we show that incorporating Twitter data improves forecasting of unemployment claims, while the original method remains marginally better at nowcasting. Enriching the input signal with temporal statistical features (e.g., moving average and rate of change) further reduces errors, and improves the predictive utility of the proposed method when applied to other economic indices, such as consumer confidence.

cs.SI↗