To use reinforcement learning from human feedback (RLHF) in practical applications, it is crucial to learn reward models from diverse sources of human feedback and to consider human factors involved in providing feedback of different types. However, the systematic study of learning from diverse types of feedback is held back by limited standardized tooling available to researchers. To bridge this gap, we propose RLHF-Blender, a configurable, interactive interface for learning from human feedback. RLHF-Blender provides a modular experimentation framework and implementation that enables researchers to systematically investigate the properties and qualities of human feedback for reward learning. The system facilitates the exploration of various feedback types, including demonstrations, rankings, comparisons, and natural language instructions, as well as studies considering the impact of human factors on their effectiveness. We discuss a set of concrete research opportunities enabled by RLHF-Blender. More information is available at https://rlhfblender.info/.

Reinforcement learning (RL) methods work in discrete time. In order to apply RL to inherently continuous problems like robotic control, a specific time discretization needs to be defined. This is a choice between sparse time control, which may be easier to train, and finer time control, which may allow for better ultimate performance. In this work, we propose SusACER, an off-policy RL algorithm that combines the advantages of different time discretization settings. Initially, it operates with sparse time discretization and gradually switches to a fine one. We analyze the effects of the changing time discretization in robotic control environments: Ant, HalfCheetah, Hopper, and Walker2D. In all cases our proposed algorithm outperforms state of the art.

Advanced video technologies are driving the development of the futuristic Metaverse, which aims to connect users from anywhere and anytime. As such, the use cases for users will be much more diverse, leading to a mix of 360-degree videos with two types: non-VR and VR 360-degree videos. This paper presents a novel Quality of Service model for heterogeneous 360-degree videos with different requirements for frame rates and cybersickness. We propose a frame-slotted structure and conduct frame-wise optimization using self-designed differentiated deep reinforcement learning algorithms. Specifically, we design two structures, Separate Input Differentiated Output (SIDO) and Merged Input Differentiated Output (MIDO), for this heterogeneous scenario. We also conduct comprehensive experiments to demonstrate their effectiveness.

Monoclonal antibodies (mAbs) have emerged as indispensable assets in medicine, and are currently at the forefront of biopharmaceutical product development. However, the growing market demand and the substantial doses required for mAb clinical treatments necessitate significant progress in its large-scale production. Most of the processes for industrial mAb production rely on batch operations, which result in significant downtime. The shift towards a fully continuous and integrated manufacturing process holds the potential to boost product yield and quality, while eliminating the extra expenses associated with storing intermediate products. The integrated continuous mAb production process can be divided into the upstream and downstream processes. One crucial aspect that ensures the continuity of the integrated process is the switching of the capture columns, which are typically chromatography columns operated in a fed-batch manner downstream. Due to the discrete nature of the switching operation, advanced process control algorithms such as economic MPC (EMPC) are computationally difficult to implement. This is because an integer nonlinear program (INLP) needs to be solved online at each sampling time. This paper introduces two computationally-efficient approaches for EMPC implementation, namely, a sigmoid function approximation approach and a rectified linear unit (ReLU) approximation approach. It also explores the application of deep reinforcement learning (DRL). These three methods are compared to the traditional switching approach which is based on a 1% product breakthrough rule and which involves no optimization.

We demonstrate equivalence between the reinforcement learning problem and the supervised classification problem. We consequently equate the exploration exploitation trade-off in reinforcement learning to the dataset imbalance problem in supervised classification, and find similarities in how they are addressed. From our analysis of the aforementioned problems we derive a novel loss function for reinforcement learning and supervised classification. Scope Loss, our new loss function, adjusts gradients to prevent performance losses from over-exploitation and dataset imbalances, without the need for any tuning. We test Scope Loss against SOTA loss functions over a basket of benchmark reinforcement learning tasks and a skewed classification dataset, and show that Scope Loss outperforms other loss functions.

Abstract: Balance assessment during physical rehabilitation often relies on rubricoriented battery tests to score a patient's physical capabilities, leading to subjectivity. While some objective balance assessments exist, they are often limited to tracking the center of pressure (COP), which does not fully capture the whole-body postural stability. This study explores the use of the center of mass (COM) state space and presents a promising avenue for monitoring the balance capabilities in humans. We employ a musculoskeletal model integrated with a balance controller, trained through reinforcement learning (RL), to investigate balancing capabilities. The RL framework consists of two interconnected neural networks governing balance recovery and muscle coordination respectively, trained using Proximal Policy Optimization (PPO) with reference state initialization, early termination, and multiple training strategies. By exploring recovery from random initial COM states (position and velocity) space for a trained controller, we obtain the final BR enclosing successful balance recovery trajectories. Comparing the BRs with analytical postural stability limits from a linear inverted pendulum model, we observe a similar trend in successful COM states but more limited ranges in the recoverable areas. We further investigate the effect of muscle weakness and neural excitation delay on the BRs, revealing reduced balancing capability in different regions. Overall, our approach of learning muscular balance controllers presents a promising new method for establishing balance recovery limits and objectively assessing balance capability in bipedal systems, particularly in humans. Keywords: Balance, Reinforcement Learning, Musculoskeletal Modeling, Bipedal Systems, Motor Disorders 1. Introduction Falls and subsequent injuries pose a significant health risk for the elderly and mobility-impaired populations. Poor balancing capabilities are the leading cause of falls in the elderly population, which reduces the overall quality of life of aging patients [1-3]. The injuries sustained by these patients can range from lower-body fractures, particularly in the hip, to head injuries, with falls being the leading cause of traumatic brain injuries [4]. Therefore, effective balance assessment and rehabilitation are critical components not only to health monitoring and injury prevention in mobility-impaired individuals, but also to the diagnoses of other serious underlying medical conditions. Since balance is maintained through a complicated network of physiological systems in the body, it is difficult to pinpoint a single origin causing deficiencies in patients and to assess balance through simple isolated measures. In most clinical environments, balance assessment is performed as a battery of balance exercises designed to evaluate the patient's ability to perform selected tasks.

The Space-Air-Ground Integrated Network (SAGIN), integrating heterogeneous devices including low earth orbit (LEO) satellites, unmanned aerial vehicles (UAVs), and ground users (GUs), holds significant promise for advancing smart city applications. However, resource management of the SAGIN is a challenge requiring urgent study in that inappropriate resource management will cause poor data transmission, and hence affect the services in smart cities. In this paper, we develop a comprehensive SAGIN system that encompasses five distinct communication links and propose an efficient cooperative multi-type multi-agent deep reinforcement learning (CMT-MARL) method to address the resource management issue. The experimental results highlight the efficacy of the proposed CMT-MARL, as evidenced by key performance indicators such as the overall transmission rate and transmission success rate. These results underscore the potential value and feasibility of future implementation of the SAGIN.

QDax is an open-source library with a streamlined and modular API for Quality-Diversity (QD) optimization algorithms in Jax. The library serves as a versatile tool for optimization purposes, ranging from black-box optimization to continuous control. QDax offers implementations of popular QD, Neuroevolution, and Reinforcement Learning (RL) algorithms, supported by various examples. All the implementations can be just-in-time compiled with Jax, facilitating efficient execution across multiple accelerators, including GPUs and TPUs. These implementations effectively demonstrate the framework's flexibility and user-friendliness, easing experimentation for research purposes. Furthermore, the library is thoroughly documented and tested with 95\% coverage.

Transportation agencies need to assess ramp metering performance when deploying or expanding a ramp metering system. The evaluation of a ramp metering strategy is primarily centered around examining its impact on freeway traffic mobility. One way these effects can be explored is by comparing traffic states, such as the speed before and after the ramp metering strategy has been altered. Predicting freeway traffic states for the after scenarios following the implementation of a new ramp metering control strategy could offer valuable insights into the potential effectiveness of the target strategy. However, the use of machine learning methods in predicting the freeway traffic state for the after scenarios and evaluating the effectiveness of transportation policies or traffic control strategies such as ramp metering is somewhat limited in the current literature. To bridge the research gap, this study presents a framework for predicting freeway traffic parameters (speed, occupancy, and flow rate) for the after situations when a new ramp metering control strategy is implemented. By learning the association between the spatial-temporal features of traffic states in before and after situations for known freeway segments, the proposed framework can transfer this learning to predict the traffic parameters for new freeway segments. The proposed framework is built upon a transfer learning model. Experimental results show that the proposed framework is feasible for use as an alternative for predicting freeway traffic parameters to proactively evaluate ramp metering performance.

To develop the most appropriate control strategy and monitor, maintain, and evaluate the traffic performance of the freeway weaving areas, state and local Departments of Transportation need to have access to traffic flows at each pair of on-ramp and off-ramp. However, ramp flows are not always readily available to transportation agencies and little effort has been made to estimate these missing flows in locations where no physical sensors are installed. To bridge this research gap, a data-driven framework is proposed that can accurately estimate the missing ramp flows by solely using data collected from loop detectors on freeway mainlines. The proposed framework employs a transfer learning model. The transfer learning model relaxes the assumption that the underlying data distributions of the source and target domains must be the same. Therefore, the proposed framework can guarantee high-accuracy estimation of on-ramp and off-ramp flows on freeways with different traffic patterns, distributions, and characteristics. Based on the experimental results, the flow estimation mean absolute errors range between 23.90 veh/h to 40.85 veh/h for on-ramps, and 31.58 veh/h to 45.31 veh/h for off-ramps; the flow estimation root mean square errors range between 34.55 veh/h to 57.77 veh/h for on-ramps, and 41.75 veh/h to 58.80 veh/h for off-ramps. Further, the comparison analysis shows that the proposed framework outperforms other conventional machine learning models. The estimated ramp flows based on the proposed method can help transportation agencies to enhance the operations of their ramp control strategies for locations where physical sensors are not installed.