mlsanityData to 14 Oct 2026

DualCast: A Dual-Path Language Model for Bimodal Financial Time-Series Forecasting

arXiv 2609.38197 ↗First posted 18 Sep 2026cs.LG cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

Financial time-series forecasting must capture price dynamics across heterogeneous assets while incorporating news available at prediction time. We introduce DualCast, a dual-path framework that extends a frozen language model with a discrete financial vocabulary. Each log-return patch is represented by a learned summary token and three residual shape tokens, preserving local drift and volatility while allowing shape patterns to be shared across assets. To improve codebook utilization, we develop adaptive frequency-equalizing residual vector quantization, which rebalances overloaded codewords without compromising reconstruction accuracy. The fast path trains only the new financial-token embeddings and output heads on a frozen Qwen3-8B backbone. A toggleable LoRA adapter enables a slow path that conditions on the fast forecast and news available at the forecast origin to produce a revised prediction. The reviser is initialized by supervised fine-tuning and further optimized with a return-space group relative policy optimization objective that rewards improvements over the fast forecast. In zero-shot evaluations covering equities and energy prices at five-minute, daily, and weekly resolutions, the slow path achieves the lowest mean absolute percentage error among the compared methods in 8 of 12 dataset-horizon settings, including every longest-horizon setting. News ablations indicate additional gains in most tested settings, although their magnitude varies across markets. DualCast thus combines a fast numerical forecaster with an optional text-conditioned revision mechanism.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 33 of its 36 references that are in the map. First posted on arXiv 18 Sep 2026.

Builds on

33 of 36 references.

References · 33

  1. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  2. 2025Qwen3 Technical Reportcited by 8,820
  3. 2017Neural Discrete Representation Learningcited by 4,712
  4. 2025DAPO: An Open-Source LLM Reinforcement Learning System at Scalecited by 2,810
  5. 2022A Time Series is Worth 64 Words: Long-term Forecasting with Transformerscited by 2,039
  6. 2023BloombergGPT: A Large Language Model for Financecited by 939
  7. 2021SoundStream: An End-to-End Neural Audio Codeccited by 771
  8. 2024Chronos: Learning the Language of Time Seriescited by 759
  9. 2023Time-LLM: Time Series Forecasting by Reprogramming Large Language Modelscited by 700
  10. 2023A decoder-only foundation model for time-series forecastingcited by 692
  11. 2023One Fits All:Power General Time Series Analysis by Pretrained LMcited by 632
  12. 2024Unified Training of Universal Time Series Forecasting Transformerscited by 614
21 more
  1. 2023Large Language Models Are Zero-Shot Time Series Forecasterscited by 500
  2. 2024MOMENT: A Family of Open Time-series Foundation Modelscited by 482
  3. 2019FinBERT: Financial Sentiment Analysis with Pre-trained Language Modelscited by 385
  4. 2024Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Expertscited by 287
  5. 2023FinGPT: Open-Source Financial Large Language Modelscited by 253
  6. 2024Timer: Generative Pre-trained Transformers Are Large Time Series Modelscited by 208
  7. 2025Chronos-2: From Univariate to Universal Forecastingcited by 194
  8. 2025Sundial: A Family of Highly Capable Time Series Foundation Modelscited by 180
  9. 2023TEST: Text Prototype Aligned Embedding to Activate LLM's Ability for Time Seriescited by 169
  10. 2024ChatTime: A Unified Multimodal Time Series Foundation Model Bridging Numerical and Textual Datacited by 100
  11. 2024Time-MMD: Multi-Domain Multimodal Dataset for Time Series Analysiscited by 74
  12. 2025MTBench: A Multimodal Time Series Benchmark for Temporal Reasoning and Question Answeringcited by 55
  13. 2024Restructuring Vector Quantization with the Rotation Trickcited by 51
  14. 2024TOTEM: TOkenized Time Series EMbeddings for General Time Series Analysiscited by 39
  15. 2023Learning to Embed Time Series Patches Independentlycited by 35
  16. 2023MASTER: Market-Guided Stock Transformer for Stock Price Forecastingcited by 32
  17. 2025Can Slow-thinking LLMs Reason Over Time? Empirical Studies in Time Series Forecastingcited by 20
  18. 2025Achieving Time Series Reasoning Requires Rethinking Model Design, Tasks Formulation, and Evaluationcited by 17
  19. 2025FinMultiTime: A Four-Modal Bilingual Dataset for Financial Time-Series Analysiscited by 11
  20. 2026From Knowing to Doing: A Memory-Controlled Benchmark for LLM Trading Agents on Stock Marketscited by 11
  21. 2025Eliciting Chain-of-Thought Reasoning for Time Series Analysis using Reinforcement Learningcited by 5

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025FinZero: Launching Multi-modal Financial Time Series Forecast with Large Reasoning Model
  2. 2025Multimodal Language Models with Modality-Specific Experts for Financial Forecasting from Interleaved Sequences of Text and Time Series
  3. 2026MM-FinEval: A Multi-Task Multimodal Benchmark for Real-World Financial Forecasting
  4. 2026RAVEN: A Regime-Aware Variable-context Expert Network for Financial Time Series Forecasting
  5. 2026HAN-Mamba: Hierarchical Selective State Space Networks for Multi-Scale Financial Volatility Forecasting
  6. 2025VISTA: Vision-Language Inference for Training-Free Stock Time-Series Analysis
  7. 2026Reasoning through Verifiable Forecast Actions: Consistency-Grounded RL for Financial LLMs
  8. 2026FinAbstain: Uncertainty-Calibrated Multimodal RAG for Selective Financial Forecasting