Domain Aggregation Networks for Multi-Source Domain Adaptation

In many real-world applications, we want to exploit multiple source datasets of similar tasks to learn a model for a different but related target dataset -- e.g., recognizing characters of a new font using a set of different fonts. While most recent research has considered ad-hoc combination rules to address this problem, we extend previous work on domain discrepancy minimization to develop a finite-sample generalization bound, and accordingly propose a theoretically justified optimization procedure. The algorithm we develop, Domain AggRegation Network (DARN), is able to effectively adjust the weight of each source domain during training to ensure relevant domains are given more importance for adaptation. We evaluate the proposed method on real-world sentiment analysis and digit recognition datasets and show that DARN can significantly outperform the state-of-the-art alternatives.

Striving for Simplicity in Off-policy Deep Reinforcement Learning

Reflecting on the advances of off-policy deep reinforcement learning (RL) algorithms since the development of DQN in 2013, it is important to ask: are the complexities of recent off-policy methods really necessary? In an attempt to isolate the contributions of various factors of variation in off-policy deep RL and to help design simpler algorithms, this paper investigates a set of related questions: First, can effective policies be learned given only access to logged offline experience? Second, how much of the benefits of recent distributional RL algorithms is attributed to improvements in exploration versus exploitation behavior? Third, can simpler off-policy RL algorithms outperform distributional RL without learning explicit distributions over returns? This paper uses a batch RL experimental setup on Atari 2600 games to investigate these questions. Unexpectedly, we find that batch RL algorithms trained solely on logged experiences of a DQN agent are able to significantly outperform online DQN. Our experiments suggest that the benefits of distributional RL mainly stem from better exploitation. We present a simple and novel variant of ensemble Q-learning called Random Ensemble Mixture (REM), which enforces optimal Bellman consistency on random convex combinations of the Q-heads of a multi-head Q-network. The batch REM agent trained offline on DQN data outperforms the batch QR-DQN and online C51 algorithms.

Advantage Amplification in Slowly Evolving Latent-State Environments

Latent-state environments with long horizons, such as those faced by recommender systems, pose significant challenges for reinforcement learning (RL). In this work, we identify and analyze several key hurdles for RL in such environments, including belief state error and small action advantage. We develop a general principle of advantage amplification that can overcome these hurdles through the use of temporal abstraction. We propose several aggregation methods and prove they induce amplification in certain settings. We also bound the loss in optimality incurred by our methods in environments where latent state evolves slowly and demonstrate their performance empirically in a stylized user-modeling task.

Exponential Family Estimation via Adversarial Dynamics Embedding

We present an efficient algorithm for maximum likelihood estimation (MLE) of the general exponential family, even in cases when the energy function is represented by a deep neural network. We consider the primal-dual view of the MLE for the kinectics augmented model, which naturally introduces an adversarial dual sampler. The sampler will be represented by a novel neural network architectures, dynamics embeddings, mimicking the dynamical-based samplers, e.g., Hamiltonian Monte-Carlo and its variants. The dynamics embedding parametrization inherits the flexibility from HMC, and provides tractable entropy estimation of the augmented model. Meanwhile, it couples the adversarial dual samplers with the primal model, reducing memory and sample complexity. We further show that several existing estimators, including contrastive divergence (Hinton, 2002), score matching (Hyv\"arinen, 2005), pseudo-likelihood (Besag, 1975), noise-contrastive estimation (Gutmann and Hyv\"arinen, 2010), non-local contrastive objectives (Vickrey et al., 2010), and minimum probability flow (Sohl-Dickstein et al., 2011), can be recast as the special cases of the proposed method with different prefixed dual samplers. Finally, we empirically demonstrate the superiority of the proposed estimator against existing state-of-the-art methods on synthetic and real-world benchmarks.

Learning to Generalize from Sparse and Underspecified Rewards

We consider the problem of learning from sparse and underspecified rewards, where an agent receives a complex input, such as a natural language instruction, and needs to generate a complex response, such as an action sequence, while only receiving binary success-failure feedback. Such success-failure rewards are often underspecified: they do not distinguish between purposeful and accidental success. Generalization from underspecified rewards hinges on discounting spurious trajectories that attain accidental success, while learning from sparse feedback requires effective exploration. We address exploration by using a mode covering direction of KL divergence to collect a diverse set of successful trajectories, followed by a mode seeking KL divergence to train a robust policy. We propose Meta Reward Learning (MeRL) to construct an auxiliary reward function that provides more refined feedback for learning. The parameters of the auxiliary reward function are optimized with respect to the validation performance of a trained policy. The MeRL approach outperforms our alternative reward learning technique based on Bayesian Optimization, and achieves the state-of-the-art on weakly-supervised semantic parsing. It improves previous work by 1.2% and 2.4% on WikiTableQuestions and WikiSQL datasets respectively.

The Value Function Polytope in Reinforcement Learning

We establish geometric and topological properties of the space of value functions in finite state-action Markov decision processes. Our main contribution is the characterization of the nature of its shape: a general polytope (Aigner et al., 2010). To demonstrate this result, we exhibit several properties of the structural relationship between policies and value functions including the line theorem, which shows that the value functions of policies constrained on all but one state describe a line segment. Finally, we use this novel perspective to introduce visualizations to enhance the understanding of the dynamics of reinforcement learning algorithms.

A Geometric Perspective on Optimal Representations for Reinforcement Learning

This paper proposes a new approach to representation learning based on geometric properties of the space of value functions. We study a two-part approximation of the value function: a nonlinear map from states to vectors, or representation, followed by a linear map from vectors to values. Our formulation considers adapting the representation to minimize the (linear) approximation of the value function of all stationary policies for a given environment. We show that this optimization reduces to making accurate predictions regarding a special class of value functions which we call adversarial value functions (AVFs). We argue that these AVFs make excellent auxiliary tasks, and use them to construct a loss which can be efficiently minimized to find a near-optimal representation for reinforcement learning. We highlight characteristics of the method in a series of experiments on the four-room domain.

Understanding the impact of entropy on policy optimization

Entropy regularization is commonly used to improve policy optimization in reinforcement learning. It is believed to help with exploration by encouraging the selection of more stochastic policies. In this work, we analyze this claim and, through new visualizations of the optimization landscape, we observe that incorporating entropy in policy optimization serves as a regularizer. We show that even with access to the exact gradient, policy optimization is difficult due to the geometry of the objective function. We qualitatively show that, in some environments, entropy regularization can make the optimization landscape smoother, thereby connecting local optima and enabling the use of larger learning rates. This manuscript presents new tools for understanding the underlying optimization landscape and highlights the challenge of designing general-purpose policy optimization algorithms in reinforcement learning.