{"id":"ddaa9c40-201a-4deb-8a14-13fa60c1572b","arxiv_id":"2610.02706","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"AdaTempo aligns demonstrations by phase, averages their relative fast-and-slow patterns, and resamples the training data so robot policies run faster with only modest success loss.","lead":"This paper presents a training-time method that rewrites slow robot demonstrations so imitation-learned policies finish tasks up to roughly 3.5 times faster, without any runtime speed-switching module. It matters for deploying learned manipulation in time-critical settings, because the retiming pattern is extracted for free from the demonstrations themselves.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek/deepseek-v4-flash","headline":"The success-rate evidence validating Eq. (7) is unquantified and, on the precision-critical Insertion task, indistinguishable from noise; the central trade-off claim is not yet statistically established.","rationale":"The reader's stated weakest_assumption (Eq. (7) monotonicity: relative tempo as a proxy for acceleration tolerance) is real but is only load-bearing in the sense that its sole validation channel — end-to-end success rate — is statistically ungrounded. That merges the reader's concern (i) (no error bars) with the Eq. (7) bridge: the bridge can only be confirmed by success rates, and those rates are within noise precisely on Insertion, the task designed to exercise the precision-cautious phases the cue claims to protect. I disagree with the reader's weighting only in emphasis: the authors already acknowledge the bound-selection overfitting and the 'empirical cue, not direct estimate' caveat ([b175], [b178]), so those are acknowledged limitations rather than new findings; what remains unresolved is whether the success-rate evidence for the trade-off claim is real at all. The real-world aggregate is heavily dependent on a single task (Conveyor Fast, +27 of +35 total successful-trial gains), and the Sorting task is a reversal. Because the mechanism itself is independently supported by the leave-one-out reproducibility analysis and the ablations, and because the flaws are fixable with variance reporting and aggregate sensitivity checks, the correct disposition is CONDITIONAL, matching the reader; no verdict change is warranted, only reinforcement of the condition. REJECT would be premature since reproducibility and ablations provide genuine support, and ACCEPT is not justified while the headline comparisons may be within noise.","tokens_in":13977,"tokens_out":4130,"duration_ms":30394,"concrete_test":"Recompute Tables II and III from per-seed success counts and report Wilson 95% (or bootstrapped binomial) intervals per task and method; state explicitly whether the Insertion 25.00%-vs-23.33% gap and the real-world Sorting 14/30-vs-15/30-vs-18/30 differences are resolvable above noise. Then recompute the aggregate real-world success rates (77.8% vs 58.3%) with Conveyor Fast removed, and report whether the claimed aggregate trade-off survives without the single dominant task. Publish these results as a supplementary table so the per-task reversals and the aggregate dependency are transparent.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that shared relative tempo yields a stronger success–speed trade-off rests on Eq. (7), which maps consensus relative tempo m(p)q(p) to per-phase acceleration. The only end-to-end check of this bridge is success rate, and that evidence is unquantified and, on the decisive task, within noise. On Insertion — the precision task where the tempo cue should matter most — AdaTempo achieves 25.00% vs ACT 23.33% and DemoSpeedup 18.00% (Table II); with 3 seeds × 100 rollouts, a 25%-vs-23.33% gap (roughly 75 vs 70 of 300) is about one standard error, i.e., compatible with chance. In the real world, Sorting is a reversal: AdaTempo 14/30 vs ACT 15/30 and DemoSpeedup 18/30 (Table III), again within noise at n=30. No table reports per-seed variance, confidence intervals, or tests, so several headline differences may be noise. The aggregate real-world gain (77.8% vs 58.3%) is dominated by the single Conveyor Fast task (+27 successful trials of a +35 total gain), and the per-task (c_min, c_max) bounds are selected on held-out validation rollouts (Sec. IV-A(c); authors acknowledge this in [b178]), so the reported 3.57x is conditional on task-specific tuning. These issues do not invalidate the reproducible-tempo mechanism (Fig. 5, leave-one-out correlations 0.861–0.952), but they fail to establish that the tempo cue beats baselines above noise.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek/deepseek-v4-flash","summary":"The paper introduces AdaTempo, an offline, self-supervised method that accelerates visuomotor policies trained by imitation learning. Demonstrations are temporally aligned by a progress-regularized monotonic DTW to a canonical medoid (Sec. III-A), converted into episode-internal relative-tempo profiles (Eq. 4), and aggregated into a consensus with MAD-based confidence (Eqs. 5-6). The consensus is mapped to a continuous speedup profile via Eq. (7), shaped by lookahead soft-minimum and asymmetric slew-rate limiting (Eqs. 8-9), and used to resample the demonstrations into an accelerated training set (Eqs. 10-11). Standard ACT and Diffusion Policy models are then trained on the resampled trajectories and deployed with no runtime tempo mechanism. Evaluation is conducted on five simulated tasks (ALOHA and robomimic) and six real-world tasks, against the original policies, uniform-acceleration baselines, DemoSpeedup, and SAIL. The paper reports up to 3.57x speedup, a mean simulation success of 75.5% at 2.83x SOD, and an aggregate real-world success gain of 77.8% versus 58.3% for ACT.","tokens_in":14468,"tokens_out":3449,"duration_ms":23771,"significance":"If the central claim holds, AdaTempo is a genuinely attractive method: it needs no auxiliary model, no runtime tempo selection, and no interaction-based optimization, and it produces a continuous, phase-dependent speedup profile from demonstrations alone. The manuscript has real strengths: the pipeline is fully specified; the reported arithmetic is internally consistent (the mean success of 75.5% and mean SOD of 2.83x reproduce from Table II, and the 2.02x/1.95x real-world speedups reproduce from Table III); and the leave-one-out consistency analysis (Fig. 5) is a genuine out-of-sample check showing that the relative-tempo cue is reproducible across demonstrations and that DTW alignment improves it on all 11 tasks (correlations 0.661-0.885 arc-length only, 0.861-0.952 DTW-aligned). The load-bearing gap is that the link from reproducible cue to acceleration tolerance, encoded by Eq. (7), is validated only by success rates, and those are reported without variance, confidence intervals, or tests, with the decisive precision task within noise.","major_comments":[{"comment":"The central success-speed trade-off claim is not statistically established. Success rates are means over only three seeds and no per-seed variance, confidence interval, or significance test is reported anywhere. On Insertion, the precision-critical task where the tempo cue should matter most, AdaTempo achieves 25.00% versus 23.33% for ACT and 18.00% for DemoSpeedup; over 300 rollouts that gap is about one standard error, i.e., compatible with chance. The sentence in Sec. IV-B that AdaTempo 'achieves the highest success on both Transfer Cube and Insertion' is therefore not supported at the task level. Please report per-seed results and confidence intervals for at least the headline tasks, and either add rollouts on Insertion or temper the claim.","section":"Sec. IV-B, Table II","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The real-world aggregate gain is dominated by a single task and at least one task is a reversal within noise. Aggregating Table III, AdaTempo's +35 total successful trials relative to ACT consist of +27 from Conveyor Fast alone (29/30 vs 2/30); on Sorting, AdaTempo is worse than DemoSpeedup (14/30 vs 18/30) and no better than ACT (15/30), all within noise at n=30. The statement that AdaTempo 'achieves the best or tied-best success on four of the six tasks' should be accompanied by uncertainty bounds; at 30 trials per condition the binomial standard error is roughly 9 percentage points, which swallows several of the reported per-task differences.","section":"Sec. IV-C, Table III","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Eq. (7) is an assumption, not a derived result: consensus relative tempo m(p)q(p), which measures how much faster or slower demonstrators move in each phase relative to their own average, is asserted to be monotonically informative about how much that phase tolerates acceleration. The leave-one-out analysis in Sec. IV-D validates the reproducibility of the cue, which is a different property. The only end-to-end check of the Eq. (7) bridge is success rate, which -- per the two comments above -- is not yet distinguished from noise on the decisive tasks. Additionally, the absolute speedup level depends on the per-task bounds (c_min, c_max) selected with held-out validation rollouts; the authors acknowledge this in the conclusion ([b178]), but as a consequence the headline 3.57x is a tuned quantity rather than a parameter-free prediction. A concrete test would be to vary (c_min, c_max) on Insertion and show a monotone success degradation consistent with the phase-wise profile, or to report success at the same SOD for AdaTempo and DemoSpeedup.","section":"Sec. III-B(b) and Sec. IV-A(c), Eq. (7)","authors_acknowledge":"Relative tempo is an empirical cue rather than a direct es- timate of the maximum feasible speed.","authors_acknowledge_p":0.64}],"minor_comments":[{"comment":"SOD is computed over successful rollouts only, with failures excluded. Since slower rollouts are more likely to fail for an accelerated policy, this metric is potentially favorable to the accelerated method; please state explicitly whether the same exclusion rule is applied to all baselines and, ideally, report SOD computed over all rollouts as a sensitivity check.","section":"Sec. IV-A(d)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The 'up to 3.57x speedup' headline is the Can task value from Table II and is conditional on the per-task (c_min, c_max) selection; consider stating in the abstract that the per-task bounds are selected with held-out validation rollouts so that readers are not misled about the strength of the prediction.","section":"Abstract and Sec. I","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The hyperparameter list is complete, which is commendable, but the sentence reporting eta_i as '0.01 mediant,si di(t,s)' is hard to parse; a short clause explaining that this scales the penalty relative to the typical local cost would help reproducibility.","section":"Sec. IV-A(c)","authors_acknowledge":null,"authors_acknowledge_p":null}],"recommendation":"major_revision"},"author_rebuttal":null,"desk_editor":{"model":"deepseek/deepseek-v4-flash","letter":"The thing to know: AdaTempo is a clean, well-specified training-side method, and the relative-tempo cue is genuinely distinct from DemoSpeedup, ESPADA, AutoSpeed, and SAIL. Episode-internal rank normalization, median consensus with MAD-based confidence, and virtual-time resampling hang together, and the leave-one-out Spearman analysis (0.861–0.952 with DTW vs 0.662–0.885 arc-length-only) is a real out-of-sample check that shared tempo structure exists. The ablations are coherent: consensus helps, confidence weighting helps, temporal regularization helps on Insertion. The arithmetic reproduces from the tables. The authors also state the main limitations themselves — the cue is empirical, not a feasibility estimate, and the bounds are selected per task on validation rollouts — so those are known conditions, not hidden flaws.\n\nThe soft spot, and it is the load-bearing one: the end-to-end claim that AdaTempo yields a stronger success–speed trade-off is not statistically established. No table reports per-seed variance or confidence intervals. On Insertion, the precision-critical task where the cue should matter most, AdaTempo's 25.00% vs ACT's 23.33% is about one standard error at 3 seeds × 100 rollouts — compatible with chance. Real-world Sorting is a reversal (14/30 vs DemoSpeedup's 18/30), again within noise at n=30. The aggregate real-world gain is dominated by Conveyor Fast (+27 successful trials of a +35 total). The conditional verdict is right, and the stress-test concern lands: Eq. (7), mapping consensus tempo to per-phase acceleration, is validated only by success rate, and that evidence is currently too weak to distinguish the mechanism from noise.\n\nTwo lesser issues. The per-task (c_min, c_max) tuning sets the absolute speedup, so \"up to 3.57x\" is conditional on task-specific validation selection; the authors acknowledge this, but the framing still overstates. And no code or data are released, which matters for a method paper whose value is largely in the pipeline.\n\nWho this is for: anyone in the demonstration-speedup subfield, and people deploying imitation-learned policies in time-critical settings. It deserves a serious referee. The right outcome is major revision: variance reporting or confidence intervals on the headline comparisons, either automatic bound selection or a sensitivity analysis showing the result is not tuned into place, and released code and data. I would cite it for the cue and the consistency analysis regardless.","headline":"A clean, well-specified training-side speedup method with a genuinely new tempo cue, but the headline success-rate advantage over baselines is not yet above noise.","tokens_in":15192,"tokens_out":2774,"duration_ms":33665},"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek/deepseek-v4-flash","headline":"AdaTempo claims demonstrations encode shared, phase-dependent tempo, and that turning the consensus into a continuous speedup profile for offline resampling yields faster, more reliable policies — up to $3.57\\times$ speedup.","keywords":["robot manipulation","imitation learning","trajectory acceleration","relative tempo","dynamic time warping","offline resampling","visuomotor policy","success-speed trade-off"],"falsifier":"A direct per-phase compression test would settle the premise: on a task with a clear cautious phase such as Insertion, compress only the cautious grasp-and-insert phase versus only the fast transport phase and compare success; if compressing the cautious phase degrades success no more than compressing the fast phase, the coupling of tempo to acceleration tolerance in the speedup mapping is wrong. A cheaper check is to record which phases failures occur in during AdaTempo rollouts — if failures concentrate in phases the consensus marked as fast and therefore safely compressible, the central assumption is falsified.","tokens_in":15392,"feed_emoji":"🤖","tokens_out":8120,"duration_ms":52752,"temperature":0.7,"pith_summary":"This paper tries to establish that teleoperated demonstrations of a manipulation task contain a reusable, self-supervised timing cue: within each episode, some phases are consistently executed faster and others (grasping, insertion) more slowly, and this relative-tempo pattern is shared across demonstrators. AdaTempo extracts that pattern by aligning demonstrations to a canonical phase via dynamic time warping, ranking each episode's phase-local tempo to strip out absolute speed, and aggregating the ranked profiles into a consensus weighted by disagreement-based confidence, which it maps to a continuous speedup profile. It then resamples the demonstrations offline by that profile and trains standard policies (action-chunk transformers such as ACT and the Diffusion Policy backbone) on the accelerated trajectories, embedding the tempo so that deployment needs no runtime module or online retiming. If the claim holds, demonstration structure alone can decide both how much and where to accelerate, and the paper reports up to $3.57\\times$ speedup over demonstrations, the highest average simulation success ($75.5\\%$ at a mean $2.83\\times$ speedup-over-demonstration), and a real-world aggregate success rise from $58.3\\%$ to $77.8\\%$ while completing tasks $2.02\\times$ faster on average. A sympathetic reader would care because this is an offline, training-side acceleration that adds no extra machinery at execution time, unlike prior adaptive methods.","feed_headline":"Common demo tempo speeds robot policies 3.57x","feed_subtitle":"AdaTempo reads which phases are naturally slow and compresses only those, beating uniform-speedup baselines.","key_machinery":"The load-bearing object is the consensus relative-tempo profile $m(p)q(p)$: the across-demonstration median of ranked, phase-local tempo ($m$) weighted by a MAD-derived confidence ($q$) evaluated at a canonical phase $p\\in[0,1]$. The episode-internal ranking of inverse phase occupancy does the essential work — normalizing tempo by rank rather than absolute speed isolates how tempo varies within a demonstration instead of how fast the whole episode runs. This profile is converted to a continuous speedup $c_i(t) = c_{\\min} + (c_{\\max}-c_{\\min})\\,m(p_i(t))\\,q(p_i(t))$, regularized by a lookahead soft minimum that anticipates cautious phases and asymmetric slew limits that let the profile slow down faster than it speeds up, then applied through offline virtual-time resampling to produce the accelerated training dataset. Phase correspondence itself is established by progress-regularized monotonic DTW to a canonical medoid, using normalized cumulative motion arc length as the progress descriptor that prevents implausible matches between distant task stages.","core_discovery":"The central claim is that cross-demonstration relative tempo is a reproducible and informative cue for phase-dependent speedup. Concretely, AdaTempo shows that within each demonstration, faster phases occupy fewer timesteps, so ranking canonical-phase bins by inverse occupancy yields an episode-relative tempo that removes each demonstrator's global speed scale; aligning demonstrations with progress-regularized dynamic time warping to a medoid makes these profiles comparable at matched task progress, and agreement with the leave-one-out consensus reaches Spearman correlations of $0.861$–$0.952$ across all eleven tasks, improving on arc-length-only correspondence in every task. The paper then treats the consensus median $m(p)$ modulated by a median-absolute-deviation-based confidence $q(p)$ as a continuous speedup profile $c_i(t) = c_{\\min} + (c_{\\max}-c_{\\min})\\,m(p_i(t))\\,q(p_i(t))$, smooths it with a lookahead soft minimum and asymmetric slew-rate limiting, and resamples each demonstration by advancing virtual source time so that denser samples fall in cautious phases. Training standard visuomotor policies on the resampled trajectories yields the reported success–speed trade-off: the highest average simulation success at a mean speedup comparable to the closest offline baseline, and the best or tied-best real-world success on four of six tasks.","pith_inferences":["The MAD-derived confidence $q(p)$ could serve as a phase-wise risk estimate on its own: phases where demonstrations disagree about tempo are plausibly the risky ones to compress, so the same consensus could gate acceleration limits without any reinforcement learning or simulator.","Because the acceleration bounds $(c_{\\min}, c_{\\max})$ are chosen per task from validation rollouts, the reported speedup level partly depends on that selection; a natural test is whether bounds can be predicted from demonstration statistics (overall speed, gripper-dwell indicators of precision) to make the pipeline fully automatic, a direction the paper flags as future work.","The paper states that a single medoid and consensus may fail when demonstrations contain qualitatively different strategies; if so, applying the same machinery within strategy clusters should raise confidence and tempo accuracy, which predicts a testable gain from strategy-aware clustering.","AdaTempo accelerates at training time, so it is arguably complementary to runtime adaptive methods: an offline schedule embedded in the policy plus a lightweight runtime feasibility check could combine the reliability reported here with the online adaptation that a fixed learned tempo cannot provide."],"forward_implications":["Uniform speedup is a suboptimal lever: because phases tolerate acceleration differently, the phase-aware consensus reaches a mean $2.83\\times$ SOD at $75.5\\%$ success in simulation, versus DemoSpeedup's comparable $2.86\\times$ at a lower $68.1\\%$.","The tempo schedule is embedded in training data, so deployment is standard policy inference with no tempo-selection module or online retiming; only low-level control gains are raised for tracking.","The approach transfers across policy backbones, improving or matching baselines on both ACT-based ALOHA tasks and Diffusion-Policy-based robomimic tasks.","Over the real-world suite, AdaTempo raises aggregate success from $58.3\\%$ (ACT) to $77.8\\%$ at a $2.02\\times$ mean completion-time speedup, with the largest reliability gain on the time-critical Conveyor Fast task ($29/30$ versus $2/30$ for ACT).","The relative-tempo structure is reproducible across demonstrators, with pairwise Spearman correlations of $0.983$, $0.941$, and $0.955$ on Conveyor Fast, supporting its use as a self-supervised supervisory signal."],"supporting_citations":[{"why":"Supplies the dynamic programming DTW alignment procedure used to establish phase correspondence between demonstrations and the canonical medoid.","marker":"[32]"},{"why":"Provides the motor-control prior that skilled movements retain relative timing under changes in execution speed, motivating the relative-tempo cue.","marker":"[12]"},{"why":"DemoSpeedup is the closest offline demonstration-resampling baseline that AdaTempo must beat on the success–speed trade-off, and the source of the RBD-style augmentation comparator.","marker":"[7]"},{"why":"SAIL is the representative runtime adaptive acceleration baseline whose reported DP numbers are quoted for comparison.","marker":"[8]"},{"why":"Supplies the ALOHA MuJoCo simulation tasks and their 50-demonstration ACT datasets used in the evaluation.","marker":"[1]"},{"why":"Supplies the robomimic Can, Lift, and Square tasks and their human demonstration datasets for the Diffusion Policy experiments.","marker":"[37]"},{"why":"Supplies the generative Diffusion Policy backbone trained on the resampled accelerated trajectories for the robomimic tasks.","marker":"[2]"}],"fun_headline_variants":["Demo tempo unlocks 3.57x faster robot manipulation","AdaTempo finds shared demo tempo, speeds policies 3.57x","Robot policies learn faster from relative demo tempo","AdaTempo compresses slow phases, hits 3.57x speedup","Shared tempo in demos speeds robot policies 3.57x"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The load-bearing premise is that the consensus relative tempo — how much slower or faster each phase is executed relative to a demonstration's own average — tracks how much that phase can be safely accelerated; reproducibility across demonstrators does not by itself establish that a cautious phase can actually be compressed, and the only end-to-end check is success rate, which is weak precisely on the precision-critical Insertion task ($25.00\\%$ versus a $23.33\\%$ baseline).","fun_headline_variants_meta":{"raw":{"variants":["Demo tempo unlocks 3.57x faster robot manipulation","AdaTempo finds shared demo tempo, speeds policies 3.57x","Robot policies learn faster from relative demo tempo","AdaTempo compresses slow phases, hits 3.57x speedup","Shared tempo in demos speeds robot policies 3.57x"]},"model":"deepseek/deepseek-v4-flash","effort":"low","cost_usd":0.000251,"raw_usage":{"cost":0.00025088,"is_byok":false,"cost_details":{"upstream_inference_cost":0.00025088,"upstream_inference_prompt_cost":0.000161,"upstream_inference_completions_cost":8.988e-05},"total_tokens":1471,"prompt_tokens":1150,"completion_tokens":321,"prompt_tokens_details":{"audio_tokens":0,"video_tokens":0,"cached_tokens":0,"cache_write_tokens":0},"completion_tokens_details":{"audio_tokens":0,"image_tokens":0,"reasoning_tokens":246}},"tokens_in":1150,"tokens_out":321,"cost_source":"openrouter_usage","duration_ms":4279,"image_pages":[],"temperature":1.0,"billed_attempts":1,"reasoning_tokens":246,"transport_provider":"openrouter","cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-10-05T19:32:53.108720+00:00","model_set":{"reader":"deepseek/deepseek-v4-flash"},"pipeline_version":"pith-pipeline@v1.4.1-daily-deepseek-or-medium","input_kind":null,"input_chars":null,"input_total_chars":null,"passes_missing":null,"published_standing":null,"published_confidence":null,"objection_audit":null,"run_record":null,"basis":{"kind":"not_recorded","source":null,"chars_given":null,"chars_total":null},"receipt":null,"falsifier":"A direct per-phase compression test would settle the premise: on a task with a clear cautious phase such as Insertion, compress only the cautious grasp-and-insert phase versus only the fast transport phase and compare success; if compressing the cautious phase degrades success no more than compressing the fast phase, the coupling of tempo to acceleration tolerance in the speedup mapping is wrong. A cheaper check is to record which phases failures occur in during AdaTempo rollouts — if failures concentrate in phases the consensus marked as fast and therefore safely compressible, the central assumption is falsified.","supporting_citations":[{"cited_title":"Dynamic programming algorithm optimiza- tion for spoken word recognition,","cited_arxiv_id":null,"evidence_quote":"Supplies the dynamic programming DTW alignment procedure used to establish phase correspondence between demonstrations and the canonical medoid."},{"cited_title":"The search for invariance in skilled movement behavior,","cited_arxiv_id":null,"evidence_quote":"Provides the motor-control prior that skilled movements retain relative timing under changes in execution speed, motivating the relative-tempo cue."},{"cited_title":"DemoSpeedup: Accelerating visuomotor policies via entropy-guided demonstration acceleration,","cited_arxiv_id":null,"evidence_quote":"DemoSpeedup is the closest offline demonstration-resampling baseline that AdaTempo must beat on the success–speed trade-off, and the source of the RBD-style augmentation comparator."},{"cited_title":"SAIL: Faster-than-demonstration execution of imitation learning policies,","cited_arxiv_id":null,"evidence_quote":"SAIL is the representative runtime adaptive acceleration baseline whose reported DP numbers are quoted for comparison."},{"cited_title":"Learning fine-grained bimanual manipulation with low-cost hardware,","cited_arxiv_id":null,"evidence_quote":"Supplies the ALOHA MuJoCo simulation tasks and their 50-demonstration ACT datasets used in the evaluation."},{"cited_title":"What matters in learning from offline human demonstrations for robot manipula- tion,","cited_arxiv_id":null,"evidence_quote":"Supplies the robomimic Can, Lift, and Square tasks and their human demonstration datasets for the Diffusion Policy experiments."},{"cited_title":"Diffusion policy: Visuomotor policy learning via action diffusion,","cited_arxiv_id":null,"evidence_quote":"Supplies the generative Diffusion Policy backbone trained on the resampled accelerated trajectories for the robomimic tasks."}],"lane_accuracy":{"status":"not yet measured","lane":{"model":"deepseek/deepseek-v4-flash","input_kind":null},"url":"https://pith.science/accuracy"},"review_version":1}