Ensemble Successor Representations for Task Generalization in Offline-to-Online Reinforcement Learning

In Reinforcement Learning (RL), training a policy from scratch with online experiences can be inefficient because of the difficulties in exploration. Recently, offline RL provides a promising solution by giving an initialized offline policy, which can be refined through online interactions. However, existing approaches primarily perform offline and online learning in the same task, without considering the task generalization problem in offline-to-online adaptation. In real-world applications, it is common that we only have an offline dataset from a specific task while aiming for fast online-adaptation for several tasks. To address this problem, our work builds upon the investigation of successor representations for task generalization in online RL and extends the framework to incorporate offline-to-online learning. We demonstrate that the conventional paradigm using successor features cannot effectively utilize offline data and improve the performance for the new task by online fine-tuning. To mitigate this, we introduce a novel methodology that leverages offline data to acquire an ensemble of successor representations and subsequently constructs ensemble Q functions. This approach enables robust representation learning from datasets with different coverage and facilitates fast adaption of Q functions towards new tasks during the online fine-tuning phase. Extensive empirical evaluations provide compelling evidence showcasing the superior performance of our method in generalizing to diverse or even unseen tasks.

Diverse Randomized Value Functions: A Provably Pessimistic Approach for Offline Reinforcement Learning

Offline Reinforcement Learning (RL) faces distributional shift and unreliable value estimation, especially for out-of-distribution (OOD) actions. To address this, existing uncertainty-based methods penalize the value function with uncertainty quantification and demand numerous ensemble networks, posing computational challenges and suboptimal outcomes. In this paper, we introduce a novel strategy employing diverse randomized value functions to estimate the posterior distribution of $Q$-values. It provides robust uncertainty quantification and estimates lower confidence bounds (LCB) of $Q$-values. By applying moderate value penalties for OOD actions, our method fosters a provably pessimistic approach. We also emphasize on diversity within randomized value functions and enhance efficiency by introducing a diversity regularization method, reducing the requisite number of networks. These modules lead to reliable value estimation and efficient policy learning from offline data. Theoretical analysis shows that our method recovers the provably efficient LCB-penalty under linear MDP assumptions. Extensive empirical results also demonstrate that our proposed method significantly outperforms baseline methods in terms of performance and parametric efficiency.

Regularized Conditional Diffusion Model for Multi-Task Preference Alignment

Sequential decision-making is desired to align with human intents and exhibit versatility across various tasks. Previous methods formulate it as a conditional generation process, utilizing return-conditioned diffusion models to directly model trajectory distributions. Nevertheless, the return-conditioned paradigm relies on pre-defined reward functions, facing challenges when applied in multi-task settings characterized by varying reward functions (versatility) and showing limited controllability concerning human preferences (alignment). In this work, we adopt multi-task preferences as a unified condition for both single- and multi-task decision-making, and propose preference representations aligned with preference labels. The learned representations are used to guide the conditional generation process of diffusion models, and we introduce an auxiliary objective to maximize the mutual information between representations and corresponding generated trajectories, improving alignment between trajectories and preferences. Extensive experiments in D4RL and Meta-World demonstrate that our method presents favorable performance in single- and multi-task scenarios, and exhibits superior alignment with preferences.

Structure-informed Positional Encoding for Music Generation

Music generated by deep learning methods often suffers from a lack of coherence and long-term organization. Yet, multi-scale hierarchical structure is a distinctive feature of music signals. To leverage this information, we propose a structure-informed positional encoding framework for music generation with Transformers. We design three variants in terms of absolute, relative and non-stationary positional information. We comprehensively test them on two symbolic music generation tasks: next-timestep prediction and accompaniment generation. As a comparison, we choose multiple baselines from the literature and demonstrate the merits of our methods using several musically-motivated evaluation metrics. In particular, our methods improve the melodic and structural consistency of the generated pieces.

Transfer Learning and Bias Correction with Pre-trained Audio Embeddings

Deep neural network models have become the dominant approach to a large variety of tasks within music information retrieval (MIR). These models generally require large amounts of (annotated) training data to achieve high accuracy. Because not all applications in MIR have sufficient quantities of training data, it is becoming increasingly common to transfer models across domains. This approach allows representations derived for one task to be applied to another, and can result in high accuracy with less stringent training data requirements for the downstream task. However, the properties of pre-trained audio embeddings are not fully understood. Specifically, and unlike traditionally engineered features, the representations extracted from pre-trained deep networks may embed and propagate biases from the model's training regime. This work investigates the phenomenon of bias propagation in the context of pre-trained audio representations for the task of instrument recognition. We first demonstrate that three different pre-trained representations (VGGish, OpenL3, and YAMNet) exhibit comparable performance when constrained to a single dataset, but differ in their ability to generalize across datasets (OpenMIC and IRMAS). We then investigate dataset identity and genre distribution as potential sources of bias. Finally, we propose and evaluate post-processing countermeasures to mitigate the effects of bias, and improve generalization across datasets.

Mesostructures: Beyond Spectrogram Loss in Differentiable Time-Frequency Analysis

Computer musicians refer to mesostructures as the intermediate levels of articulation between the microstructure of waveshapes and the macrostructure of musical forms. Examples of mesostructures include melody, arpeggios, syncopation, polyphonic grouping, and textural contrast. Despite their central role in musical expression, they have received limited attention in deep learning. Currently, autoencoders and neural audio synthesizers are only trained and evaluated at the scale of microstructure: i.e., local amplitude variations up to 100 milliseconds or so. In this paper, we formulate and address the problem of mesostructural audio modeling via a composition of a differentiable arpeggiator and time-frequency scattering. We empirically demonstrate that time--frequency scattering serves as a differentiable model of similarity between synthesis parameters that govern mesostructure. By exposing the sensitivity of short-time spectral distances to time alignment, we motivate the need for a time-invariant and multiscale differentiable time--frequency model of similarity at the level of both local spectra and spectrotemporal modulations.

Differentiable Time-Frequency Scattering in Kymatio

Joint time-frequency scattering (JTFS) is a convolutional operator in the time-frequency domain which extracts spectrotemporal modulations at various rates and scales. It offers an idealized model of spectrotemporal receptive fields (STRF) in the primary auditory cortex, and thus may serve as a biological plausible surrogate for human perceptual judgments at the scale of isolated audio events. Yet, prior implementations of JTFS and STRF have remained outside of the standard toolkit of perceptual similarity measures and evaluation methods for audio generation. We trace this issue down to three limitations: differentiability, speed, and flexibility. In this paper, we present an implementation of time-frequency scattering in Kymatio, an open-source Python package for scattering transforms. Unlike prior implementations, Kymatio accommodates NumPy and PyTorch as backends and is thus portable on both CPU and GPU. We demonstrate the usefulness of JTFS in Kymatio via three applications: unsupervised manifold learning of spectrotemporal modulations, supervised classification of musical instruments, and texture resynthesis of bioacoustic sounds.

Joint Scattering for Automatic Chick Call Recognition

Animal vocalisations contain important information about health, emotional state, and behaviour, thus can be potentially used for animal welfare monitoring. Motivated by the spectro-temporal patterns of chick calls in the time$-$frequency domain, in this paper we propose an automatic system for chick call recognition using the joint time$-$frequency scattering transform (JTFS). Taking full-length recordings as input, the system first extracts chick call candidates by an onset detector and silence removal. After computing their JTFS features, a support vector machine classifier groups each candidate into different chick call types. Evaluating on a dataset comprising 3013 chick calls collected in laboratory conditions, the proposed recognition system using the JTFS features improves the frame- and event-based macro F-measures by 9.5% and 11.7%, respectively, than that of a mel-frequency cepstral coefficients baseline.

Impact Mitigation for Dynamic Legged Robots with Steel Wire Transmission Using Nonlinear Active Compliance Control

Impact mitigation is crucial to the stable locomotion of legged robots, especially in high-speed dynamic locomotion. This paper presents a leg locomotion system including the nonlinear active compliance control and the active impedance control for the steel wire transmission-based legged robot. The developed control system enables high-speed dynamic locomotion with excellent impact mitigation and leg position tracking performance, where three strategies are applied. a) The feed-forward controller is designed according to the linear motor-leg model with the information of Coulomb friction and viscous friction. b) Steel wire transmission model-based compensation guarantees ideal virtual spring compliance characteristics. c) Nonlinear active compliance control and active impedance control ensure better impact mitigation performance than linear scheme and guarantee position tracking performance. The proposed control system is verified on a real robot named SCIT Dog and the experiment demonstrates the ideal impact mitigation ability in high-speed dynamic locomotion without any passive spring mechanism.

Continuous-time Gaussian Process Trajectory Generation for Multi-robot Formation via Probabilistic Inference

In this paper, we extend a famous motion planning approach GPMP2 to multi-robot cases, yielding a novel centralized trajectory generation method for the multi-robot formation. A sparse Gaussian Process model is employed to represent the continuous-time trajectories of all robots as a limited number of states, which improves computational efficiency due to the sparsity. We add constraints to guarantee collision avoidance between individuals as well as formation maintenance, then all constraints and kinematics are formulated on a factor graph. By introducing a global planner, our proposed method can generate trajectories efficiently for a team of robots which have to get through a width-varying area by adaptive formation change. Finally, we provide the implementation of an incremental replanning algorithm to demonstrate the online operation potential of our proposed framework. The experiments in simulation and real world illustrate the feasibility, efficiency and scalability of our approach.