United States

Many applications of reinforcement learning can be formalized as goal-conditioned environments, where, in each episode, there is a &quot;goal&quot; that affects the rewards obtained during that episode but does not affect the dynamics. Various techniques have been proposed to improve performance in goal-conditioned environments, such as automatic curriculum generation and goal relabeling. In this work, we explore a connection between off-policy reinforcement learning in goal-conditioned settings and knowledge distillation. In particular: the current Q-value function and the target Q-value estimate are both functions of the goal, and we would like to train the Q-value function to match its target for all goals. We therefore apply Gradient-Based Attention Transfer (Zagoruyko and Komodakis 2017), a knowledge distillation technique, to the Q-function update. We empirically show that this can improve the performance of goal-conditioned off-policy reinforcement learning when the space of goals is high-dimensional. We also show that this technique can be adapted to allow for efficient learning in the case of multiple simultaneous sparse goals, where the agent can attain a reward by achieving any one of a large set of objectives, all specified at test time. Finally, to provide theoretical support, we give examples of classes of environments where (under some assumptions) standard off-policy algorithms such as DDPG require at least O(d^2) replay buffer transitions to learn an optimal policy, while our proposed technique requires only O(d) transitions, where d is the dimensionality of the goal and state space. Code and appendix are available at https://github.com/alevine0/ReenGAGE. 

AAAI 2023

Goal-Conditioned Q-Learning as Knowledge Distillation

goal-conditioned

knowledge distillation

reinforcement learning

The purpose of the AAAI conference series is to promote research in Artificial Intelligence (AI) and foster scientific exchange between researchers, practitioners, scientists, students, and engineers across the entirety of AI and its affiliated disciplines. AAAI-23 is the Thirty-Seventh AAAI Conference on Artificial Intelligence. The theme of this conference is to create collaborative bridges within and beyond AI. Like previous AAAI conferences, AAAI-23 will feature technical paper presentations, special tracks, invited speakers, workshops, tutorials, poster sessions, senior member presentations, competitions, and exhibit programs, and two new activities: a Bridge Program and a Lab Program. Many of these activities are tailored to the theme of bridges and all are selected according to the highest standards, with additional programs for students and young researchers. 
AAAI is providing you with a conference planner, which you can use to help organize your itinerary of activities. This includes talks to attend in person, talks to attend remotely, breaks with colleagues and your site seeing activities. To access this conference planner, please go to [https://aaai-2023.takemobi.io](https://aaai-2023.takemobi.io).

In order to access this site, you need to register. If you haven't already, please register [here](https://aaai.org/Conferences/AAAI-23/registration/).


The purpose of the AAAI conference series is to promote research in Artificial Intelligence (AI) and foster scientific exchange between researchers, practitioners, scientists, students, and engineers across the entirety of AI and its affiliated disciplines.

Many applications of reinforcement learning can be formalized as goal-conditioned environments, where, in each episode, there is a "goal" that affects the rewards obtained during that episode but does not affect the dynamics. Various techniques have been proposed to improve performance in goal-conditioned environments, such as automatic curriculum generation and goal relabeling. In this work, we explore a connection between off-policy reinforcement learning in goal-conditioned settings and knowledge distillation. In particular: the current Q-value function and the target Q-value estimate are both functions of the goal, and we would like to train the Q-value function to match its target for all goals. We therefore apply Gradient-Based Attention Transfer (Zagoruyko and Komodakis 2017), a knowledge distillation technique, to the Q-function update. We empirically show that this can improve the performance of goal-conditioned off-policy reinforcement learning when the space of goals is high-dimensional. We also show that this technique can be adapted to allow for efficient learning in the case of multiple simultaneous sparse goals, where the agent can attain a reward by achieving any one of a large set of objectives, all specified at test time. Finally, to provide theoretical support, we give examples of classes of environments where (under some assumptions) standard off-policy algorithms such as DDPG require at least O(d^2) replay buffer transitions to learn an optimal policy, while our proposed technique requires only O(d) transitions, where d is the dimensionality of the goal and state space. Code and appendix are available at https://github.com/alevine0/ReenGAGE. 

poster

Tabular biomedical data is often high-dimensional but with a very small number of samples. Although recent work showed that well-regularised simple neural networks could outperform more sophisticated architectures on tabular data, they are still prone to overfitting on tiny datasets with many potentially irrelevant features. To combat these issues, we propose Weight Predictor Network with Feature Selection (WPFS) for learning neural networks from high-dimensional and small sample data by reducing the number of learnable parameters and simultaneously performing feature selection. In addition to the classification network, WPFS uses two small auxiliary networks that together output the weights of the first layer of the classification model. We evaluate on nine real-world biomedical datasets and demonstrate that WPFS outperforms other standard as well as more recent methods typically applied to tabular data. Furthermore, we investigate the proposed feature selection mechanism and show that it improves performance while providing useful insights into the learning task.

Weight Predictor Network with Feature Selection for Small Sample Tabular Biomedical Data

Subgraph isomorphism counting is an important problem on graphs, as many graph-based tasks exploit recurring subgraph patterns. Classical methods usually boil down to a backtracking framework that needs to navigate a huge search space with prohibitive computational cost. Some recent studies resort to graph neural networks (GNNs) to learn a low-dimensional representation for both the query and input graphs, in order to predict the number of subgraph isomorphisms on the input graph. However, typical GNNs employ a node-centric message passing scheme that receives and aggregates messages on nodes, which is inadequate in complex structure matching for isomorphism counting. Moreover, on an input graph, the space of possible query graphs is enormous, and different parts of the input graph will be triggered to match different queries. Thus, expecting a fixed representation of the input graph to match diversely structured query graphs is unrealistic. In this paper, we propose a novel GNN called Count-GNN for subgraph isomorphism counting, to deal with the above challenges. At the edge level, given that an edge is an atomic unit of encoding graph structures, we propose an edge-centric message passing scheme, where messages on edges are propagated and aggregated based on the edge adjacency to preserve fine-grained structural information. At the graph level, we modulate the input graph representation conditioned on the query, so that the input graph can be adapted to each query individually to improve their matching. Finally, we conduct extensive experiments on a number of benchmark datasets to demonstrate the superior performance of Count-GNN.

Learning to Count Isomorphisms with Graph Neural Networks

Active learning (AL) aims to sample the most informative data instances for labeling, which makes the model fitting data efficient while significantly reducing the annotation cost. However, most existing AL models make a strong assumption that the annotated data instances are always assigned correct labels, which may not hold true in many practical settings.  In this paper, we develop a theoretical framework to formally analyze the impact of noisy annotations and show that systematically re-sampling guarantees to reduce the noise rate, which can lead to improved generalization capability. More importantly, the theoretical framework demonstrates the key benefit of conducting active re-sampling on label-efficient learning, which is critical for AL. The theoretical results also suggest essential properties of an active re-sampling function with a fast convergence speed and guaranteed error reduction. This inspires us to design a novel spatial-temporal active re-sampling function by leveraging the important spatial and temporal properties of maximum-margin classifiers. Extensive experiments conducted on both synthetic and real-world data clearly demonstrate the effectiveness of the proposed active re-sampling function.  

STARS: Spatial-Temporal Active Re-Sampling for Label-Efficient Learning from Noisy Annotations

Competition between traditional platforms is known to improve user utility by aligning the platform's actions with user preferences. But to what extent is alignment exhibited in data-driven marketplaces? To study this question from a theoretical perspective, we introduce a duopoly market where platform actions are bandit algorithms and the two platforms compete for user participation. A salient feature of this market is that the quality of recommendations depends on both the bandit algorithm and the amount of data provided by interactions from users. This interdependency between the algorithm performance and the actions of users complicates the structure of market equilibria and their quality in terms of user utility. Our main finding is that competition in this market does not perfectly align market outcomes with user utility. Interestingly, market outcomes exhibit misalignment not only when the platforms have separate data repositories, but also when the platforms have a shared data repository. Nonetheless, the data sharing assumptions impact what mechanism drives misalignment and also affect the specific form of misalignment (e.g. the quality of the best-case and worst-case market outcomes). More broadly, our work illustrates that competition in digital marketplaces has subtle consequences for user utility that merit further investigation. 

Competition, Alignment, and Equilibria in Digital Marketplaces

As a direct depth sensor, radar holds promise as a tool to improve monocular 3D object detection, which suffers from depth errors, due in part to the depth-scale ambiguity. On the other hand, leveraging radar depths is hampered by difficulties in precisely associating radar returns with 3D estimates from monocular methods, effectively erasing its benefits. This paper proposes a fusion network that addresses this radar-camera association challenge. We train our network to predict the 3D offsets between radar returns and object centers, enabling radar depths to enhance the accuracy of 3D monocular detection. By using parallel radar and camera backbones, our network fuses information at both the feature level and detection level, while at the same time leveraging a state-of-the-art monocular detection technique without retraining it. Experimental results show significant improvement in mean average precision and translation error on the nuScenes dataset over monocular counterparts. Our source code is available at https://github.com/longyunf/radiant.

RADIANT: Radar-Image Association Network for 3D Object Detection

This paper studies restless multi-armed bandit (RMAB) problems with unknown arm transition dynamics but with known correlated arm features. The goal is to learn a model to predict transition dynamics given features, where the Whittle index policy solves the RMAB problems using predicted transitions. However, prior works often learn the model by maximizing the predictive accuracy instead of final RMAB solution quality, causing a mismatch between training and evaluation objectives. To address this shortcoming, we propose a novel approach for decision-focused learning in RMAB that directly trains the predictive model to maximize the Whittle index solution quality. We present three key contributions: (i) we establish differentiability of the Whittle index policy to support decision-focused learning; (ii) we significantly improve the scalability of decision-focused learning approaches in sequential problems, specifically RMAB problems; (iii) we apply our algorithm to a previously collected  dataset of maternal and child health to demonstrate its performance. Indeed, our algorithm is the first for decision-focused learning in RMAB that scales to real-world problem sizes.

Scalable Decision-Focused Learning in Restless Multi-Armed Bandits with Application to Maternal and Child Health

The concept of potential outcome types is one of the fundamental components of causal inference. However, even in randomized experiments, assumptions on the data generating process, such as monotonicity (Tian and Pearl, 2000), are required to evaluate the probabilities of the potential outcome types. To solve the problem without such assumptions in experimental studies, a novel identification condition based on proxy covariate information is proposed in this paper. In addition, the estimation problem of the probabilities of the potential outcome types reduces to that of singular models when they are identifiable through the proposed condition. Thus, they cannot be evaluated by standard statistical estimation methods. To overcome this difficulty, new plug-in estimators of these probabilities are presented, and the asymptotic normality of the proposed estimators is shown. 

Probabilities of Potential Outcome Types in Experimental Studies: Identification and Estimation Based on Proxy Covariate Information

We formalize and analyze a fundamental component of dif-
ferentiable neural architecture search (NAS): local “opera-
tion scoring” at each operation choice. We view existing
operation scoring functions as inexact proxies for accuracy,
and we find that they perform poorly when analyzed empir-
ically on NAS benchmarks. From this perspective, we intro-
duce a novel perturbation-based zero-cost operation scor-
ing (Zero-Cost-PT) approach, which utilizes zero-cost prox-
ies that were recently studied in multi-trial NAS but de-
grade significantly on larger search spaces, typical for dif-
ferentiable NAS. We conduct a thorough empirical evalu-
ation on a number of NAS benchmarks and large search
spaces, from NAS-Bench-201, NAS-Bench-1Shot1, NAS-
Bench-Macro, to DARTS-like and MobileNet-like spaces,
showing significant improvements in both search time and
accuracy. On the ImageNet classification task on the DARTS
search space, our approach improved accuracy compared to
the best current training-free methods (TE-NAS) while be-
ing over 10× faster (total searching time 25 minutes on a
single GPU), and observed significantly better transferabil-
ity on architectures searched on the CIFAR-10 dataset with
an accuracy increase of 1.8 pp. Our code is available at:
https://github.com/zerocostptnas/zerocost operation score.

Zero-Cost Operation Scoring in Differentiable Architecture Search

Multi-arm bandit (MAB) and stochastic linear bandit (SLB) are important models in reinforcement learning, and it is well-known that classical algorithms for bandits with time horizon $T$ suffer $\Omega(\sqrt{T})$ regret. In this paper, we study MAB and SLB with quantum reward oracles and propose quantum algorithms for both models with $O(\poly(\log T))$ regrets, exponentially improving the dependence in terms of $T$. To the best of our knowledge, this is the first provable quantum speedup for regrets of bandit problems and in general exploitation in reinforcement learning. Compared to previous literature on quantum exploration algorithms for MAB and reinforcement learning, our quantum input model is simpler and only assumes quantum oracles for each individual arm.

Quantum Multi-Armed Bandits and Stochastic Linear Bandits Enjoy Logarithmic Regrets

Although increasing model size can enhance the adversarial robustness of deep neural networks, in resource-constrained environments, however, there exist critical sparsity constraints. While the recent robust pruning technologies show promising direction to obtain adversarially robust sparse networks, they perform poorly with high sparsity. In this work, we bridge this performance gap by reparameterizing network parameters to simultaneously learns the sparse structure and the robustness. Specifically, we introduce Twin-Rep, which reparameterizes original weights into the product of two factors during training and perform pruning on the reparameterized weights to satisfy the target sparsity constraint. Twin-Rep implicitly adds the sparsity constraint without changing the robust training objective, thus can enhance robustness under high sparsity. We also introduce another variant of weight reparameterization for better channel pruning. When inferring, we restore the original weight structure to obtain compact and robust networks. Extensive experiments on diverse datasets demonstrate that our method achieves state-of-the-art results, outperforming the current sparse robust training method and robustness-aware pruning method.

Premium content

Goal-Conditioned Q-Learning as Knowledge Distillation

Downloads

Next from AAAI 2023

Weight Predictor Network with Feature Selection for Small Sample Tabular Biomedical Data

Similar lecture

The Sufficiency of Off-Policyness and Soft Clipping: PPO is still Insufficient according to an Off-Policy Measure