mlsanityData to 14 Oct 2026

GALA: Generation-Aware Cross-Modal Alignment for Text-to-Time-Series Synthesis

arXiv 2608.13741 ↗First posted 13 Aug 2026cs.CL cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Synthesizing time series from natural language is emerging as the most expressive form of controllable time series generation. However, existing text-conditioned generators either take caption embeddings frozen from off-the-shelf text encoders, or adapt the encoder end-to-end, letting the denoising loss shape the embeddings only as a by-product. In either case, the conditioning representation is never deliberately matched to the signal modality, leaving it ill-suited to guide generation. We address this by introducing GALA: Generation-Aware cross-modaL Alignment for text conditional time series generation. GALA is a two-stage approach that first contrastively couples a pretrained text encoder with a time-series foundation model into a shared embedding space with both encoders adapted to generation by an auxiliary generative loss, and then freezes the resulting caption embedding to drive a flow-matching generator. On TSFragment-600K, spanning four domains and three fragment lengths, GALA sets a new state of the art, ranking first in 30 of 36 metric columns and reaching an average rank of 1.08/1.08/1.42 at lengths 24/48/96 against 1.92/2.00/1.75 for the strongest baseline. We further find that generator-internal text encoders force a trade-off between fidelity and caption adherence, whereas conditioning on the aligned embedding breaks it: FID, CTTP, and JFTSD all improve at once. Ablating the auxiliary loss degrades FID, CTTP and JFTSD together, it indicates the generative term is a necessary component of the alignment rather than an add-on.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 36 of its 44 references that are in the map. First posted on arXiv 13 Aug 2026.

Builds on

36 of 44 references.

References · 36

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  2. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
  3. 2023Visual Instruction Tuningcited by 7,869
  4. 2018Representation Learning with Contrastive Predictive Codingcited by 7,777
  5. 2022Hierarchical Text-Conditional Image Generation with CLIP Latentscited by 6,585
  6. 2022Scalable Diffusion Models with Transformerscited by 6,054
  7. 2022Flamingo: a Visual Language Model for Few-Shot Learningcited by 4,386
  8. 2021Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervisioncited by 3,864
  9. 2020DiffWave: A Versatile Diffusion Model for Audio Synthesiscited by 1,354
  10. 2022CoCa: Contrastive Captioners are Image-Text Foundation Modelscited by 1,274
  11. 2023Time-LLM: Time Series Forecasting by Reprogramming Large Language Modelscited by 700
  12. 2023C-Pack: Packed Resources For General Chinese Embeddingscited by 578
24 more
  1. 2019On Variational Bounds of Mutual Informationcited by 559
  2. 2025Chronos-2: From Univariate to Universal Forecastingcited by 194
  3. 2024Diffusion-TS: Interpretable Diffusion for General Time Series Generationcited by 184
  4. 2023TEST: Text Prototype Aligned Embedding to Activate LLM's Ability for Time Seriescited by 169
  5. 2021TimeVAE: A Variational Auto-Encoder for Multivariate Time Series Generationcited by 142
  6. 2025EmbeddingGemma: Powerful and Lightweight Text Representationscited by 139
  7. 2024ChatTS: Aligning Time Series with LLMs via Synthetic Data for Enhanced Understanding and Reasoningcited by 103
  8. 2024Time-MMD: Multi-Domain Multimodal Dataset for Time Series Analysiscited by 74
  9. 2023Vector Quantized Time Series Generation with a Bidirectional Prior Modelcited by 50
  10. 2024Utilizing Image Transforms and Diffusion Models for Generative Modeling of Short and Long Time Seriescited by 43
  11. 2024FlowTS: Time Series Generation via Rectified Flowcited by 22
  12. 2022TTS-CGAN: A Transformer Time-Series Conditional GAN for Biosignal Data Augmentationcited by 21
  13. 2025T2S: High-resolution Time Series Generation with Text-to-Series Diffusion Modelscited by 20
  14. 2025BRIDGE: Bootstrapping Text to Control Time-Series Generation via Multi-Agent Iterative Optimization and Diffusion Modelingcited by 19
  15. 2025DiffuSETS: 12-lead ECG Generation Conditioned on Clinical Text Reports and Patient-Specific Informationcited by 16
  16. 2025With Limited Data for Multimodal Alignment, Let the STRUCTURE Guide Youcited by 16
  17. 2025Time Series Generation Under Data Scarcity: A Unified Generative Modeling Approachcited by 14
  18. 2024M3-JEPA: Multimodal Alignment via Multi-gate MoE based on the Joint-Embedding Predictive Architecturecited by 13
  19. 2025Towards Cross-Modality Modeling for Time Series Analytics: A Survey in the LLM Eracited by 12
  20. 2025Time Series Language Model for Descriptive Caption Generationcited by 11
  21. 2025Teaching Time Series to See and Speak: Forecasting with Aligned Visual and Textual Perspectivescited by 7
  22. 2025BALM-TSF: Balanced Multimodal Alignment for LLM-Based Time Series Forecastingcited by 6
  23. 2025WaveStitch: Flexible and Fast Conditional Time Series Generation with Diffusion Modelscited by 4
  24. 2026ConTSG-Bench: A Unified Benchmark for Conditional Time Series Generationcited by 4

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025ChronoSteer: Bridging Large Language Model and Time Series Foundation Model via Synthetic Cross-Modal Alignment Dataset
  2. 2024Hierarchical Multimodal LLMs with Semantic Space Alignment for Enhanced Time Series Classification
  3. 2024NeST: Neighborhood-aware semantic alignment and temporal modulation for LLM based time series forecasting
  4. 2025Aurora: Towards Universal Generative Multimodal Time Series Forecasting
  5. 2025Adapting LLMs to Time Series Forecasting via Temporal Heterogeneity Modeling and Representation Alignment
  6. 2025Enhancing LLMs for Time Series Forecasting via Structure-Guided Cross-Modal Alignment
  7. 2024TS-HTFA: Advancing Time Series Forecasting via Hierarchical Text-Free Alignment with Large Language Models
  8. 2026CodeTS: Verifiable Text-to-Time Series Generation via Executable Code