Attenuated in-context identification in time-series foundation models: diagnosis under counterfactual inputs and repair by synthetic forced-system fine-tuning

Key points

• We test this on forced engineering systems with exact counterfactuals, comparing Chronos-2, TimesFM-2.5 and TabPFN-TS with classical system identification fitted to the same context.1
• Chronos-2 identifies dynamics in context but attenuates them.1
• Context dither at inference lowers the what-if error on all six synthetic classes without training.1
• Paired counterfactual inputs, together with shuffled future inputs on measured records, test two properties: whether the covariate interface can represent dynamics and whether the pretraining prior covers the plant's time scale.1

Sources (1)

• [1]Attenuated in-context identification in time-series foundation models: diagnosis under counterfactual inputs and repair by synthetic forced-system fine-tuning
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 6, 10:39 AM

  Covariate-aware time-series foundation models (TSFMs) promise training-free what-if answers for instrumented plants: the change in output that a different future input would cause.

  We test this on forced engineering systems with exact counterfactuals, comparing Chronos-2, TimesFM-2.5 and TabPFN-TS with classical system identification fitted to the same context.

Extractive summary: sentences quoted from the sources.