It is taken as initial evidence for market time series as an input modality in financial text classification on the task of classifying sentences from Federal Reserve communication as hawkish, dovish, or neutral.
Abstract
Financial text is produced and interpreted within a market environment, yet financial text classifiers almost always receive text alone. We study whether financial time series are useful as an additional input on the task of classifying sentences from Federal Reserve communication as hawkish, dovish, or neutral. Our system, \lfts{}, extends the \lf{} architecture with this modality: a small voting network combines three independently trained components, a fine-tuned RoBERTa encoder, a prompted large language model (LLM), and a fused ensemble of time-series transformers over the market series of the months preceding publication. Because only about a thousand annotated sentences are available for training, the RoBERTa encoder is first pre-trained on sentences annotated automatically by the LLM and only then fine-tuned on the human labels. Trained on Federal Open Market Committee (FOMC) communication up to 2015 and evaluated on 2015--2022, the fused system achieves 70.2\% weighted F1 -- against 64.1\% for the zero-shot LLM -- and overtakes it with as few as 240 human-labelled sentences. We take this as initial evidence for market time series as an input modality in financial text classification.
Financial sentiment classifiers are commonly evaluated against human labels, but strong linguistic performance does not necessarily imply economically useful return predictability. This study separates these questions through two experiments. First, we construct a unified three-class benchmark from five financial text datasets and compare TF--IDF Naive Bayes, off-the-shelf FinBERT and Financial-RoBERTa encoders, zero-shot Qwen2.5-7B, and QLoRA-adapted Qwen2.5-7B, LLaMA3-8B, and Mistral-7B models. Mistral-7B achieves the best test accuracy (0.8840) and macro-F1 (0.8771), while QLoRA raises Qwen2.5's macro-F1 from 0.7274 to 0.8615. An inverse-frequency class-weighted loss does not improve Qwen2.5. Second, we evaluate economic validity on a temporally separate 2019 Benzinga sample containing 10,637 unique headlines and 13,115 headline--stock observations for a fixed S\&P~100 universe. Model probabilities are converted into continuous sentiment scores, aggregated by stock and signal date, and aligned with next-session returns over one-, two-, three-, and five-day horizons. All seven downstream models produce positive but small mean rank information coefficients at the one-day horizon; the largest is 0.0143 for FinBERT. None of the 28 model--horizon tests remains significant after Newey--West inference and false-discovery-rate correction. Portfolio results likewise fail to establish a robust advantage for the best-performing classifiers. The findings show that QLoRA is effective for financial sentiment adaptation, while also documenting a clear gap between classification accuracy and tradable cross-sectional signals.
Financial sentiment analysis has long relied on labeled data to fine-tune models like FinBERT, a process that is both costly and time-consuming. The arrival of large language models (LLMs) has changed the landscape: with zero-shot and few-shot prompting, one can now extract sentiment from financial texts using few or no annotated examples. This survey takes stock of how LLMs are being applied to this task. It begins by clarifying the core ideas behind in-context learning and chain-of-thought prompting. It then examines a range of prompt designs that have been developed to cope with the peculiarities of financial writing, such as numerical expressions, implicit sentiment and long documents. A comparison of LLM performance on standard benchmarks against fine-tuned domain models shows that general-purpose LLMs are often competitive, especially when prompts are carefully crafted. Yet three problems remain unresolved: numerical reasoning errors, hallucination, and the practical hurdles of cost, latency, and privacy. These challenges are discussed in detail, and retrieval-augmented generation, trustworthiness frameworks, and efficient open-source models are pointed out as the most promising paths forward. Overall, LLMs offer a flexible and annotation-light alternative to traditional fine-tuning, but their successful deployment in finance will depend on robust prompt engineering and solid factual grounding.
Liyang Gao· Applied and Computational En...· 0 citations
Injecting frozen relational-encoder embeddings as soft tokens into a large language model (LLM) is a conceptually appealing fusion strategy: the encoder handles multi-table structure, the LLM handles language and reasoning, and no lossy text serialization is required. We test this hypothesis concretely by injecting embeddings from a frozen Relational Transformer (RT) into Qwen3.5-4B via a learned MLP projection and LoRA adaptation, trained first with supervised fine-tuning (SFT) on chain-of-thought reasoning traces and then with group-based reinforcement learning (GSPO). We evaluate across 10 binary classification tasks on 6 relational databases from RelBench, under four supervision regimes: single-task (ST), within-dataset (WD), cross-dataset (CD), and all-task (ALL). The hybrid model does not consistently outperform standalone RT: it is frequently below random, highly sensitive to serialization format and relational-token budget, and unstable under RL training. We report these negative results and analyze the failure modes, arguing that soft-token fusion requires stronger alignment objectives and schema-aware design before it can serve as a reliable route to relational prediction.
Francisco Galuppo Azevedo, Clarissa Lima Loures· 0 citations
This study deploys a scalable machine learning pipeline: combining a transformer-based classifier applied to 2.01 million English-language AI-related news headlines (July 2022–July 2024) with large-language-model and human-annotator validation (three annotators, Fleiss’ κ=0.80) on stratified subsamples, to extract six interpretable, bias-linked discourse indicators computed at the AI-domain level: evaluative orientation (valence), loss salience, narrative drift, exposure-adjusted sentiment, cross-source divergence, and novelty-phase framing. Each operationalizes an established cognitive-psychology construct as a computable property of the information environment associated with biased risk–benefit reasoning. Results show systematic variation across domains: technical and methodological areas such as deep learning and natural language processing exhibit gain-salient framing, while safety-critical topics such as deepfakes (loss-to-gain headline ratio = 3.17) and facial recognition show strongly loss-salient profiles. Cross-model validation using an LLM on a stratified sample of 1000 headlines confirms that domain-level indicator rankings are robust to classifier choice (Spearman ρ=0.83; p<0.001), establishing the rank stability of pipeline outputs independently of the specific classification architecture. As a contextual application, domain-level profiles are mapped to European Union AI governance instruments, documenting parallels between discourse patterns and regulatory risk tiers. The framework provides a scalable, reproducible methodology for monitoring evaluative conditions in technology news across domains, sources, and time.
O. Topal, Inna Novalija, Joao Pita Costa et al.· Applied Informatics· 0 citations
Multimodal time-series forecasting has emerged as a promising paradigm in which natural-language context is expected to improve predictive performance. Recent multimodal foundation models, including Aurora, as well as early- and late-fusion approaches such as MM-TSFlib and TaTS, report substantial gains over unimodal baselines on the Time-MMD benchmark, attributing these improvements to textual information. However, whether these models are actually sensitive to the semantic content of the text remains unverified. We address this question through controlled text perturbations, attribution analyses, and probes of Aurora's text pathway. On Time-MMD, swapping each row's text for any other real text (empty, constant, within-domain shuffled, or cross-domain) moves mean MSE by less than $0.5\%$ on all three architectures. The improvement reported in the literature is recovered when a co-shipped numeric column is removed without touching text. We conclude that, on this benchmark and within this family of frozen-encoder architectures, text content is not the operative signal behind the reported gains. To support future work on text integration in multimodal foundation models for structured data, we release our perturbation protocol and evaluation harness as a reusable diagnostic toolkit.
K. Sridhar, Atharva Gupta, Nishant Pradhan et al.· 0 citations
Many text classification decisions are viable based on constituent excerpts alone. Taking inspiration from the field of multiple instance learning, we present an algorithm for training a neural network to classify text by selecting such excerpts. We show that our approach is also scalable with demonstrated learning against samples with nearly 1M tokens. We evaluate our methods on 7 datasets with emphasis on long-textual collections that far exceed the encoding limit of our base model. We present state-of-the-art results with this algorithm on 3 datasets: identification of political bias in news outlets, trigger warnings in long stories, and demographic characteristics of authors in tweet collections. Furthermore, the model trained on weakly-labeled collections of text (bags) generalizes to accurately classify constituent, smaller instances. Besides a new state-of-the-art for these problems, this approach is one of the few neural methods to excel in these datasets.
John J. Cadigan, Dayne Freitag, Eric Yeh· 0 citations
