- Rating
- 3
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
The paper proposes an unified in-context learning framework for diverse time series tasks, including forecasting, classification, anomaly detection, etc. The framework can be utilized for a specific task with only a support set of contexts rather than fine-tuning a task-specific head. Specifically, the framework is consisted by a Hierarchical Multi-Scope Transformer Encoder for temporal and covariate information capturing and a Task-Conditioned Patch Decoder for aggregating information among support and query samples. Trained by a meta-learning strategy, the method's effectiveness is demonstrated on several benchmarks across diverse time series tasks.
Critical strengths
-
The paper is well-written and easy to follow.
-
The method is well motivated.
-
The idea of a unified framework for diverse time series tasks without fine-tuning task-specific head is interesting.
-
Experiments on several benchmarks across diverse time series tasks demonstrate the effectiveness of the proposed method.
Critical weaknesses
-
Limited comparisons and discussions with ICL-based methods. Similar ideas about use in-context learning for time series forecasting were already proposed in [1]. Moreover, the idea of ICL with within-sample self-attention and compressed cross-example interface has also been widely used in Tabular Foundation Models such as TabPFN-3 and TabICLv2. Clarifying the connections and differences with these methods would help distinguish the specific contributions of the proposed method. Also, comparing the existing TSFMs that without ICL looks not fair. The comparisons with ICL-based methods is also expected.
-
For the experiments, I'm really curious about the relationship among different training tasks. Are they mostly positively transferring the knowledge or negatively? I'd like to see some empirical results of ICL training of each single task (for example forecasting) compared with the joint one. Moreover, in table 3, ICL gains not such obviously for forecasting tasks, especially for multivariate and covariate settings. Is there any analysis? Without ICL, the method still outperforms the existing TSFMs. What's the reason? Is this brought by the diverse training tasks?
-
I'm also curious about the training and inference costs of the proposed method with ICL, compare with the baseline methods. Not only the inference time, but also the FLOPs and memory usage.
[1] In-Context Fine-Tuning for Time-Series Foundation Models
Questions for authors
-
The proposed method do the information interaction between support and query samples with only the summary tokens, which reduce the computational costs but may also introduce an information bottleneck. I'm wondering would a total self-attention between all support and query tokens be better than only do cross-sample attention of the summary tokens. And would a larger amount of summary tokens be better?
-
During training, the method includes diverse training tasks. How does the proposed method balance the training steps, losses, and gradients of different tasks?
6 comments
Continued Rebuttals (2/4): Details for W1 - relation to ICL methods and the new comparisons
Connections and differences. iAmTime specializes the compression-based ICL principle of TabICLv2 and TabPFN-3 for demonstration-conditioned temporal tasks. Semantic tokens summarize designated regions (target history, demonstrated outputs, historical and future-known covariates) through masked Token Read; [START]/[MID] tokens exchange information across examples; and Token Write injects the inferred task into the query patches through FiLM. This read–exchange–write step repeats at every encoder layer. TimesFM-ICF [1] conditions a forecaster on related series placed in its context, and the task remains forecasting. Our contribution is the structured temporal interface and its episodic multi-task training.
ICL comparisons already in the submission.
- Forecasting. Chronos-2 (cross-series conditioning) and the TabPFN-based forecasters, including TabPFN-TS-3 (labeled feature–target rows). On GIFT-Eval, iAmTime, Chronos-2 and TabPFN-TS score 0.680, 0.698 and 0.771 MASE (Table 19).
- Classification. Three ICL classifiers that predict from labeled support examples without weight updates (Table 12): iAmTime obtains 71.03% mean accuracy, versus 60.66% for TiCT, 70.10% for TimEE and 72.86% for RocketPFN. Two of the ten datasets are in-domain for iAmTime; the eight-dataset means are in our response to Reviewer tHBy (Part 4).
- Mechanism. Demonstration interventions (Table 4), the NoCEA and NoTWrite ablations (Table 5, Appendix J) and representation analyses (Appendix K) examine how the interface conveys task information.
TimesFM-ICF (forecasting). TimesFM-ICF has no public checkpoint, so we follow its rolling-window ETT protocol (Zhou et al., 2021) and compare with its published results (MAE; 512-step query history; TimesFM-ICF with 50 in-context examples, iAmTime with a random four sampled from the pool of 50 constructed using their selection procedure).
| Dataset | ETTh1 | ETTh1 | ETTh2 | ETTh2 | ETTm1 | ETTm1 | ETTm2 | ETTm2 | Avg | Avg |
|---|---|---|---|---|---|---|---|---|---|---|
| Horizon | 96 | 192 | 96 | 192 | 96 | 192 | 96 | 192 | 96 | 192 |
| TimesFM-ICF (50 examples) | 0.374 | 0.400 | 0.327 | 0.371 | 0.331 | 0.364 | 0.239 | 0.283 | 0.318 | 0.355 |
| iAmTime (4 examples) | 0.293 | 0.338 | 0.309 | 0.315 | 0.295 | 0.304 | 0.186 | 0.242 | 0.271 | 0.300 |
iAmTime with four examples is better on all eight settings, by 15% on average at both horizons.
ICTP (held-out imputation). ICTP is a reference for held-out-task transfer, so we compare on imputation and not on forecasting. iAmTime-NoImp, trained without imputation, is evaluated under ICTP's protocol (details in our response to Reviewer M6wx, Part 2). Entries are standardized MSE / MAE.
| Held-out imputation | ETTh1 | ETTm1 | Weather |
|---|---|---|---|
| TimesFM+ICTP (reported) | 0.785 / 0.592 | 0.934 / 0.692 | 0.231 / 0.279 |
| LPTM+ICTP (reported) | 1.141 / 0.779 | 1.096 / 0.764 | 0.331 / 0.411 |
| iAmTime-NoImp, 4 demos | 0.759 / 0.580 | 0.910 / 0.657 | 0.224 / 0.275 |
iAmTime-NoImp has lower MSE and MAE than both published references on all three datasets (2.6–3.3% lower MSE than TimesFM+ICTP and 17–34% lower than LPTM+ICTP).
Flat Transformer (all four task families). Comparable ICL baselines for imputation and anomaly detection are lacking, so we trained a control under the same support–query formulation: a vanilla Transformer with 12 layers, hidden size 1024 and 16 heads (~300M parameters; parameter- and budget-matched), with the same explicit role tokens, a shared quantile head and staged training. The prompt is flattened into one sequence, so query patches attend directly to the demonstrations. CB-ICL, the transfer of our mechanism to a Chronos-Bolt backbone, is from Table 21.
| Model | fev skill (%) ↑ | GIFT CRPS ↓ | UCR ICL acc. (%) ↑ | Block-8 MSE ↓ | Anomaly F1 (%) ↑ |
|---|---|---|---|---|---|
| Flat Transformer (new) | 44.5 | 0.487 | 72.5 | 0.0991 | 88.71 |
| CB-ICL, 0 demos | 41.7 | 0.534 | – | 0.3261 | 77.83 |
| CB-ICL, 4 demos | 43.2 | 0.512 | – | 0.2002 | 85.16 |
| Full iAmTime | 46.8 | 0.465 | 81.3 | 0.0727 | 93.57 |
CB-ICL classification is reported on the ten multivariate datasets of Table 12 (69.35%, against 71.03% for iAmTime). With prompt information matched, full iAmTime leads the flat Transformer on every metric, including 8.8 points higher UCR accuracy and 26.6% lower Block-8 MSE. The CB-ICL pair and the flat Transformer show that demonstration gains also transfer across backbones.
Continued Rebuttals (3/4): Details for W2 - task relationships and the role of demonstrations in forecasting
Single-task versus joint training. The forecasting-only comparison is reported in Tables 5 and 17 (NoMeta, NoExmp, Meta-NoExmp):
| Variant (Table 5) | Training tasks | Training demos | fev skill (%) ↑ | GIFT CRPS ↓ | UCR acc. ↑ | Block-8 MSE ↓ | Anomaly F1 ↑ |
|---|---|---|---|---|---|---|---|
| NoExmp | forecasting only | no | 46.4 | 0.471 | 0.238 | 0.1935 | 72.10 |
| NoMeta | forecasting only | yes | 46.4 | 0.468 | 0.224 | 0.1740 | 74.30 |
| Meta-NoExmp | all tasks | no | 47.5 | 0.458 | 0.240 | 0.0814 | 92.20 |
| Full iAmTime | all tasks | yes | 46.8 | 0.465 | 0.813 | 0.0727 | 93.57 |
- Forecasting. Joint training does not hurt forecasting. With training demonstrations absent, adding the auxiliary tasks raises fev skill from 46.4% to 47.5% and improves CRPS from 0.471 to 0.458, so diverse tasks strengthen forecasting even with empty support.
- Removing one task family (Table 5). The effect is concentrated on the removed task: NoCls lowers UCR accuracy from 81.3% to 27.4%, NoImp raises Block-8 MSE from 0.0727 to 0.1421, and NoAD lowers anomaly F1 from 93.57% to 80.64%. Effects on retained tasks are small and go either way; dropping synthetic classification, for example, improves GIFT CRPS from 0.465 to 0.460 and lowers UCR accuracy to 61.1%.
- Single-task specialist (new). Block-8 imputation, MSE / MAE, same training budget:
| Training task mix | 0 examples ↓ | 4 examples ↓ |
|---|---|---|
| NoImp (joint, imputation held out) | 0.1891 / 0.2090 | 0.1421 / 0.1784 |
| Imputation-only ICL training | 0.0982 / 0.1450 | 0.0820 / 0.1311 |
| iAmTime (full joint ICL) | 0.1730 / 0.1967 | 0.0727 / 0.1265 |
With four demonstrations, the joint model beats the specialist, which supports positive transfer. The specialist scores well at zero demonstrations because its training fixes the task. NoImp still improves with four demonstrations, showing adaptation to a held-out task.
Why demonstrations matter less for forecasting than for other tasks. In Table 3, four demonstrations improve SQL skill by 0.7 points on univariate tasks and leave multivariate and covariate-informed skill unchanged; average win rate changes by +5.6, +3.8 and −0.2 points. The query already specifies the forecasting task and supplies the history, related variates and covariates, while the default demonstrations are earlier windows of the same series (Appendix E.1) and largely repeat that information. Classification needs the examples to define the label mapping: accuracy rises from 20.6% to 81.3% when they are supplied (Table 4). Forecasting does respond to example quality: perturbing the demonstrations on held-out domains lowers win rate from 71.3% to 66.2% (Appendix J.8).
More informative demonstrations (new). We supplied related series from the same dataset, found by nearest-neighbor search over iAmTime encoder embeddings; related-series conditioning has a precedent in TimesFM-ICF. Entries are SQL skill / win rate (%).
| fev-bench subset | 0 examples | 4: same-series history (Table 3) | 4: random cross-series | 4: related series |
|---|---|---|---|---|
| Univariate | 37.10 / 75.90 | 37.80 / 81.50 | 37.85 / 81.91 | 38.35 / 82.60 |
| Multivariate | 58.20 / 74.50 | 58.20 / 78.30 | 58.10 / 77.80 | 58.85 / 79.00 |
| With covariates | 48.40 / 73.50 | 48.40 / 73.30 | 48.35 / 73.45 | 48.93 / 74.50 |
Related examples are best in every subset, improving skill by 0.5–0.7 points over the default. The gain also appears in the multivariate and covariate subsets, where the default demonstrations leave skill unchanged.
Why forecasting remains strong with zero demonstrations. The auxiliary tasks contribute (Meta-NoExmp vs NoExmp above). The larger factors are per-example fusion and the staged curriculum, whose removal costs 9.0 and 16.6 fev skill points (Table 5).
| Evaluation | Metric | Full, 0 ex | Full, 4 ex | NoMeta, 0 ex | NoMeta, 4 ex |
|---|---|---|---|---|---|
| fev-bench | SQL skill (%) ↑ | 46.6 | 46.8 | 45.2 | 46.4 |
| GIFT-Eval | CRPS ↓ | 0.467 | 0.465 | 0.477 | 0.468 |
At zero demonstrations the full model is better than forecasting-only NoMeta, which shows the benefit of multi-task training for query-only forecasting. Demonstrations add little to the full checkpoint but recover almost all of the gap for NoMeta. The paper's central claim is demonstration-specified task adaptation; we will state that the forecasting benefit of demonstrations depends on the setting, and add the related-series result.
Rebuttal by Authors
Dear Reviewer pRFK,
Thank you for pointing us to related ICL work and for asking us to test the summary bottleneck directly. We answer each point below and report additional experiments that support the paper's claims. This part gives our answers and the key results; Parts 2–4 give the full tables.
W1 — ICL-based methods (Part 2). iAmTime shares the principle of within-sample attention with a compressed cross-example interface with TabICLv2 and TabPFN-3, and we will cite this connection. The difference from these methods and from TimesFM-ICF [1] is what the examples do: in iAmTime each demonstration is an explicit input–output pair, so the demonstrations define the task for one fixed checkpoint. The submission already compares with ICL methods (Chronos-2 and TabPFN-TS/-3 for forecasting; TiCT, TimEE and RocketPFN for classification). We added three comparisons with independent demonstration-based methods:
W2 — Relationship among tasks, and the role of ICL in forecasting (Part 3). Transfer is positive overall: joint training preserves forecasting accuracy (46.8 vs 46.4 fev skill for forecasting-only training), and the joint model beats an imputation-only specialist (Block-8 MSE 0.0727 vs 0.0820). On Table 3: the query already specifies the forecasting task, so the default same-series demonstrations add little; with related-series demonstrations, skill improves in all three subsets. Strength without ICL comes from the diverse training tasks (+1.1 skill points) and, above all, from per-example fusion and the curriculum (9.0 and 16.6 points).
W3 — Cost (Part 4). On a common workload with four demonstrations, iAmTime and Chronos-2 have comparable latency (23.5 vs 23.7 ms per query); iAmTime uses 31% more FLOPs and about 0.5 GiB more memory. Four demonstrations cost 6.2× the FLOPs of none.
Q1 — Full self-attention, and more summary tokens (Part 4). We ran both, with four demonstrations.
Full attention between all support and query tokens, at the same parameter count, is worse on all five metrics at 2.4× the FLOPs. More summary tokens give small, non-monotonic gains (at most 0.7 accuracy points).
Q2 — Balancing tasks (Part 4). We use no gradient surgery or loss weighting. Balance comes from curriculum sampling probabilities, one shared quantile loss in normalized target space, and demonstration-conditioned FiLM and expert routing. Sampling controls task exposure and joint AdamW optimizes the shared loss.
We will add these results and attributions to the paper and are glad to answer any follow-up questions.