Trivium: Temporal Regret as a First-Class Objective for Causal-Memory Controllers
Quick Answer
Trivium introduces long-horizon temporal regret as a key objective for causal-memory controllers, enhancing error correction in agentic systems.
Quick Take
It demonstrates logarithmic temporal regret on CausalBench-Seq, outperforming linear growth seen in outcome-only models, with preliminary validation from real- streams.
Key Points
- Temporal regret captures the duration of miscalibration before correction in causal models.
- Outcome-only learning fails to distinguish causal from spurious structures without intervention.
- Trivium shows O(log E) temporal regret on CausalBench-Seq, contrasting with linear baselines.
- Five falsifiable predictions were pre-registered for the Trivium model.
- Self-learning involves revising causal models, not retraining LLM weights.
Paper Resources
Article Content
From source RSS / original summaryarXiv:2606. 04421v1 Announce Type: new Abstract: Many current agentic systems and pipelines correct mistakes by optimizing outcome reward. This addresses only the what of failure: when an outcome diverges from prediction, the why and when of the mismatch are not systematically logged, reviewed, or corrected, so the same error can recur episode after episode. We argue that this is a structural problem, not merely a model-capacity one.
We propose long-horizon temporal regret as a first-class objective alongside outcome regret and epistemic regret over the working causal model. Temporal regret captures when failure persists: how long a miscalibrated causal model is tolerated before correction. Epistemic regret captures why failure persists: residual uncertainty or error in the working causal model. Together, the three regrets give a falsifiable account of what, why, and when a long-lived agent can fail.
Modeling the agent as a stream of E episodes, we prove three conditional results under explicit causal-probing, persistence, and detectability assumptions. First, under observationally equivalent confounding, outcome-only learning cannot distinguish causal from spurious structure without an intervention channel, so temporal miscalibration can persist linearly even after outcome regret is driven to zero.
Second, with a persistent causal log and budgeted probes, total probe complexity is logarithmic in the episode horizon, inducing O(log E) temporal regret. Third, under K detectable change-points, the rate extends to O(K log E). We instantiate Trivium and pre-register five falsifiable predictions. On CausalBench-Seq, Trivium follows the predicted logarithmic envelope while outcome-only baselines grow linearly.
A pilot real-LLM stream provides preliminary external-validity evidence across one full E = 500 run and three E = 100 frontier-model pilots. Self-learning here means revising an external causal model, not retraining LLM weights.
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.AI
See more →HOBA: Hierarchical On-Policy Bidding Agents for Adaptive Online Advertising
HOBA (Hierarchical On-policy Bidding Agents) is a novel hierarchical reinforcement learning framework that enhances online advertising bidding systems by improving adaptability and reducing hyperparameter tuning costs. It utilizes a for hyperparameter inference, a SARSA agent for expert model selection, and a dynamic expert pool for bid execution, achieving a +3.6% increase in target cost during large-scale deployment and outperforming state-of-the-art baselines on AuctionNet.