mlsanityData to 13 Oct 2026

BRIDGE: Bootstrapping Text to Control Time-Series Generation via Multi-Agent Iterative Optimization and Diffusion Modeling

arXiv 2503.02445 ↗First posted 4 Mar 2025cs.LG cs.CL cs.MAcited by 18

On the map: Diffusion models → Flow matching, time series and science

also in 2 more
Abstract (arXiv)

Time-series Generation (TSG) is a prominent research area with broad applications in simulations, data augmentation, and counterfactual analysis. While existing methods have shown promise in unconditional single-domain TSG, real-world applications demand for cross-domain approaches capable of controlled generation tailored to domain-specific constraints and instance-level requirements. In this paper, we argue that text can provide semantic insights, domain information and instance-specific temporal patterns, to guide and improve TSG. We introduce ``Text-Controlled TSG'', a task focused on generating realistic time series by incorporating textual descriptions. To address data scarcity in this setting, we propose a novel LLM-based Multi-Agent framework that synthesizes diverse, realistic text-to-TS datasets. Furthermore, we introduce BRIDGE, a hybrid text-controlled TSG framework that integrates semantic prototypes with text description for supporting domain-level guidance. This approach achieves state-of-the-art generation fidelity on 11 of 12 datasets, and improves controllability by up to 12% on MSE and 6% MAE compared to no text input generation, highlighting its potential for generating tailored time-series data.

Key facts. Cited by 18 papers in the map (data through 13 Oct 2026). Placed in Flow matching, time series and science, one of 14 directions in Diffusion models. Builds on 49 of its 77 references that are in the map. First posted on arXiv 4 Mar 2025.

Builds on

49 of 77 references.

Closest ancestors · 7

Same direction, newest first

  1. 2024TimeLDM: Latent Diffusion Model for Unconditional Time Series Generation5 Jul
  2. 2024MG-TSD: Multi-Granularity Time Series Diffusion Models with Guided Learning Process9 Mar
  3. 2024Diffusion-TS: Interpretable Diffusion for General Time Series Generation4 Mar
  4. 2023Predict, Refine, Synthesize: Self-Guiding Diffusion Models for Probabilistic Time Series Forecasting21 Jul
  5. 2023Non-autoregressive Conditional Diffusion Models for Time Series Prediction8 Jun
  6. 2023Diffusion Models for Time Series Applications: A Survey1 May
  7. 2021CSDI: Conditional Score-based Diffusion Models for Probabilistic Time Series Imputation7 Jul

Other references · 42

  1. 2020Denoising Diffusion Probabilistic Modelscited by 20,952
  2. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,376
  3. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,502
  4. 2023Self-Refine: Iterative Refinement with Self-Feedbackcited by 3,367
  5. 2022A Time Series is Worth 64 Words: Long-term Forecasting with Transformerscited by 2,016
  6. 2022TimesNet: Temporal 2D-Variation Modeling for General Time Series Analysiscited by 1,043
  7. 2022Large Language Models Are Human-Level Prompt Engineerscited by 1,031
  8. 2019N-BEATS: Neural basis expansion analysis for interpretable time series forecastingcited by 751
  9. 2024Chronos: Learning the Language of Time Seriescited by 737
  10. 2023Time-LLM: Time Series Forecasting by Reprogramming Large Language Modelscited by 693
  11. 2023One Fits All:Power General Time Series Analysis by Pretrained LMcited by 623
  12. 2024Unified Training of Universal Time Series Forecasting Transformerscited by 602
30 more
  1. 2023Large Language Models Are Zero-Shot Time Series Forecasterscited by 497
  2. 2024Sora: A Review on Background, Technology, Limitations, and Opportunities of Large Vision Modelscited by 473
  3. 2019Financial Time Series Forecasting with Deep Learning : A Systematic Literature Review: 2005-2019cited by 273
  4. 2021Monash Time Series Forecasting Archivecited by 261
  5. 2024Foundation Models for Time Series Analysis: A Tutorial and Surveycited by 237
  6. 2022Diffusion-based Time Series Imputation and Forecasting with Structured State Space Modelscited by 195
  7. 2022PromptCast: A New Prompt-based Learning Paradigm for Time Series Forecastingcited by 172
  8. 2023TEMPO: Prompt-based Generative Pre-trained Transformer for Time Series Forecastingcited by 172
  9. 2021TimeVAE: A Variational Auto-Encoder for Multivariate Time Series Generationcited by 141
  10. 2023UniTime: A Language-Empowered Unified Model for Cross-Domain Time Series Forecastingcited by 141
  11. 2023Can ChatGPT Forecast Stock Price Movements? Return Predictability and Large Language Modelscited by 127
  12. 2023Large Language Models as Evolutionary Optimizerscited by 123
  13. 2024Language Models Still Struggle to Zero-shot Reason about Time Seriescited by 83
  14. 2024Time-MMD: Multi-Domain Multimodal Dataset for Time Series Analysiscited by 70
  15. 2023The Wall Street Neophyte: A Zero-Shot Analysis of ChatGPT Over MultiModal Stock Movement Prediction Challengescited by 59
  16. 2024LSTPrompt: Large Language Models as Zero-Shot Time Series Forecasters by Long-Short-Term Promptingcited by 58
  17. 2023On the Constrained Time-Series Generation Problemcited by 52
  18. 2023Vector Quantized Time Series Generation with a Bidirectional Prior Modelcited by 49
  19. 2023Language Models Can Improve Event Prediction by Few-Shot Abductive Reasoningcited by 44
  20. 2022GT-GAN: General Purpose Time Series Synthesis with Generative Adversarial Networkscited by 43
  21. 2023ChatGPT Informed Graph Neural Network for Stock Movement Predictioncited by 41
  22. 2024Time Weaver: A Conditional Time Series Generation Modelcited by 36
  23. 2025TimeDP: Learning to Generate Multi-Domain Time Series with Domain Promptscited by 21
  24. 2023PULSAR at MEDIQA-Sum 2023: Large Language Models Augmented by Synthetic Dialogue Convert Patient Dialogues to Medical Recordscited by 11
  25. 2024Controllable Financial Market Generation with Diffusion Guided Meta Agentcited by 7
  26. 2023Drafting Event Schemas using Language Modelscited by 6
  27. 2024Towards Controllable Time Series Generationcited by 4
  28. 2024InvDiff: Invariant Guidance for Bias Mitigation in Diffusion Modelscited by 4
  29. 2024Thread Detection and Response Generation using Transformers with Prompt Optimisationcited by 2
  30. 2022Towards Neural Numeric-To-Text Generation From Temporal Personal Health Datacited by 1

Built on by

Cited by year (2026 to 13 Oct)

Citing papers per year: 2025: 12; 2026: 6.

18 papers.

Sort:

By area:

  1. 2025MIRA: Medical Time Series Foundation Model for Real-World Health Datacited by 22
  2. 2025A Survey of Reasoning and Agentic Systems in Time Series with Large Language Modelscited by 21
  3. 2025How Can Time Series Analysis Benefit From Multiple Modalities? A Survey and Outlookcited by 20
  4. 2025LLM-Enhanced Multimodal Fusion for Cross-Domain Sequential Recommendationcited by 11
  5. 2025TarDiff: Target-Oriented Diffusion Guidance for Synthetic Electronic Health Record Time Series Generationcited by 8
  6. 2025Causal Time Series Generation via Diffusion Modelscited by 7
  7. 2025Does Acceleration Cause Hidden Instability in Vision Language Models? Uncovering Instance-Level Divergence Through a Large-Scale Empirical Studycited by 2
  8. 2026Diff-MN: Diffusion Parameterized MoE-NCDE for Continuous Time Series Generation with Irregular Observationscited by 2
  9. 2025Generating Realistic Multi-Beat ECG Signalscited by 1
  10. 2025Instruction-based Time Series Editingcited by 1
  11. 2025Privacy-Preserving Generation of Clinical Narratives from Medical Terminologiescited by 1
  12. 2025CAD-Tokenizer: Towards Text-based CAD Prototyping via Modality-Specific Tokenizationcited by 1
6 more
  1. 2026Spectral-Aware Text-to-Time Series Generation with Billion-Scale Multimodal Meteorological Datacited by 1
  2. 2025Arg-LLaDA: Argument Summarization via Large Language Diffusion Models and Sufficiency-Aware Refinementcited by 0
  3. 2026UPLOTS: A Unified Pretrained Language Model for Constrained Time-series Generationcited by 0
  4. 2026Represent, Then Generate: Multimodal-Conditioned Time-Series Generation under Irregular Missingnesscited by 0
  5. 2026GALA: Generation-Aware Cross-Modal Alignment for Text-to-Time-Series Synthesiscited by 0
  6. 2026CodeTS: Verifiable Text-to-Time Series Generation via Executable Codecited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025T2S: High-resolution Time Series Generation with Text-to-Series Diffusion Models
  2. 2025Empowering Time Series Analysis with Synthetic Data: A Survey and Outlook in the Era of Foundation Models
  3. 2024Agentic Retrieval-Augmented Generation for Time Series Analysis
  4. 2025ChronoSteer: Bridging Large Language Model and Time Series Foundation Model via Synthetic Cross-Modal Alignment Dataset
  5. 2025TS-Agent: Understanding and Reasoning Over Raw Time Series via Iterative Insight Gathering
  6. 2026KairosAgent: Agentic Time Series Forecasting with Fused Semantic Reasoning
  7. 2025FLAIRR-TS -- Forecasting LLM-Agents with Iterative Refinement and Retrieval for Time Series
  8. 2026Agents of Diffusion: Enhancing Diffusion Language Models with Multi-Agent Reinforcement Learning for Structured Data Generation (Extended Version)