Mobile video analysis systems often encounter various deploying environments, where environment shifts present greater demands for responsiveness in adaptations of deployed "expert DNN models". Existing model adaptation frameworks primarily operate in a cloud-centric way, exhibiting degraded performance during adaptation and delayed reactions to environment shifts. Instead, this paper proposes MOCHA, a novel framework optimizing the responsiveness of continuous model adaptation through hierarchical collaborations between mobile and cloud resources. Specifically, MOCHA (1) reduces adaptation response delays by performing on-device model reuse and fast fine-tuning before requesting cloud model retrieval and end-to-end retraining; (2) accelerates history expert model retrieval by organizing them into a structured taxonomy utilizing domain semantics analyzed by a cloud foundation model as indices; (3) enables efficient local model reuse by maintaining onboard expert model caches for frequent scenes, which proactively prefetch model weights from the cloud model database. Extensive evaluations with real-world videos on three DNN tasks show MOCHA improves the model accuracy during adaptation by up to 6.8% while saving the response delay and retraining time by up to 35.5x and 3.0x respectively.

Machine learning systems deployed in safety-critical robotics settings must be robust to distribution shifts. However, system designers must understand the cause of a distribution shift in order to implement the appropriate intervention or mitigation strategy and prevent system failure. In this paper, we present a novel framework for diagnosing distribution shifts in a streaming fashion by deploying multiple stochastic martingales simultaneously. We show that knowledge of the underlying cause of a distribution shift can lead to proper interventions over the lifecycle of a deployed system. Our experimental framework can easily be adapted to different types of distribution shifts, models, and datasets. We find that our method outperforms existing work on diagnosing distribution shifts in terms of speed, accuracy, and flexibility, and validate the efficiency of our model in both simulated and live hardware settings.

Evolutionary search via the quality-diversity (QD) paradigm can discover highly performing solutions in different behavioural niches, showing considerable potential in complex real-world scenarios such as evolutionary robotics. Yet most QD methods only tackle static tasks that are fixed over time, which is rarely the case in the real world. Unlike noisy environments, where the fitness of an individual changes slightly at every evaluation, dynamic environments simulate tasks where external factors at unknown and irregular intervals alter the performance of the individual with a severity that is unknown a priori. Literature on optimisation in dynamic environments is extensive, yet such environments have not been explored in the context of QD search. This paper introduces a novel and generalisable Dynamic QD methodology that aims to keep the archive of past solutions updated in the case of environment changes. Secondly, we present a novel characterisation of dynamic environments that can be easily applied to well-known benchmarks, with minor interventions to move them from a static task to a dynamic one. Our Dynamic QD intervention is applied on MAP-Elites and CMA-ME, two powerful QD algorithms, and we test the dynamic variants on different dynamic tasks.

Modern machine learning approaches excel in static settings where a large amount of i.i.d. training data are available for a given task. In a dynamic environment, though, an intelligent agent needs to be able to transfer knowledge and re-use learned components across domains. It has been argued that this may be possible through causal models, aiming to mirror the modularity of the real world in terms of independent causal mechanisms. However, the true causal structure underlying a given set of data is generally not identifiable, so it is desirable to have means to quantify differences between models (e.g., between the ground truth and an estimate), on both the observational and interventional level. In the present work, we introduce the Interventional Kullback-Leibler (IKL) divergence to quantify both structural and distributional differences between models based on a finite set of multi-environment distributions generated by interventions from the ground truth. Since we generally cannot quantify all differences between causal models for every finite set of interventional distributions, we propose a sufficient condition on the intervention targets to identify subsets of observed variables on which the models provably agree or disagree.

This paper presents a novel non-stationary dynamic pricing algorithm design, where pricing agents face incomplete demand information and market environment shifts. The agents run price experiments to learn about each product's demand curve and the profit-maximizing price, while being aware of market environment shifts to avoid high opportunity costs from offering sub-optimal prices. The proposed ACIDP extends information-directed sampling (IDS) algorithms from statistical machine learning to include microeconomic choice theory, with a novel pricing strategy auditing procedure to escape sub-optimal pricing after market environment shift. The proposed ACIDP outperforms competing bandit algorithms including Upper Confidence Bound (UCB) and Thompson sampling (TS) in a series of market environment shifts.

We present an empirical study on the use of continual learning (CL) methods in a reinforcement learning (RL) scenario, which, to the best of our knowledge, has not been described before. CL is a very active recent research topic concerned with machine learning under non-stationary data distributions. Although this naturally applies to RL, the use of dedicated CL methods is still uncommon. This may be due to the fact that CL methods often assume a decomposition of CL problems into disjoint sub-tasks of stationary distribution, that the onset of these sub-tasks is known, and that sub-tasks are non-contradictory. In this study, we perform an empirical comparison of selected CL methods in a RL problem where a physically simulated robot must follow a racetrack by vision. In order to make CL methods applicable, we restrict the RL setting and introduce non-conflicting subtasks of known onset, which are however not disjoint and whose distribution, from the learner's point of view, is still non-stationary. Our results show that dedicated CL methods can significantly improve learning when compared to the baseline technique of "experience replay".