A5053630715- Reinforcement Learning in Robotics
- Data Stream Mining Techniques
- Evolutionary Algorithms and Applications
- Multilevel Inverters and Converters
- Adversarial Robustness in Machine Learning
- Robot Manipulation and Learning
- Plant and animal studies
- Microgrid Control and Optimization
- Machine Learning and Algorithms
- Insect and Arachnid Ecology and Behavior
- Bee Products Chemical Analysis
- Adaptive Dynamic Programming Control
- Advanced Bandit Algorithms Research
- vaccines and immunoinformatics approaches
- Smart Grid Energy Management
- Multi-Agent Systems and Negotiation
- Electric Vehicles and Infrastructure
- SARS-CoV-2 and COVID-19 Research
- Transportation and Mobility Innovations
- Silicon Carbide Semiconductor Technologies
- Gene Regulatory Network Analysis
- Bioinformatics and Genomic Networks
- Teaching and Learning Programming
- Parallel Computing and Optimization Techniques
- Wind Turbine Control Systems
Lawrence Livermore National Laboratory
2021-2025
University of Michigan–Dearborn
2023
Universidade de São Paulo
2014-2020
Universidade de Brasília
2019
The University of Texas at Austin
2018
Universidade Estadual de Londrina
2015

 Multiagent Reinforcement Learning (RL) solves complex tasks that require coordination with other agents through autonomous exploration of the environment. However, learning a task from scratch is impractical due to huge sample complexity RL algorithms. For this reason, reusing knowledge can come previous experience or indispensable scale up multiagent This survey provides unifying view literature on reuse in RL. We define taxonomy solutions for general problem, providing comprehensive...
The number of Electric Vehicle (EV) owners is expected to significantly increase in the near future, since EVs are regarded as valuable assets both for transportation and energy storage purposes. However, recharging a large fleet during peak hours may overload transformers distribution grid. Although several methods have been proposed flatten peak-hour loads recharge fairly possible available time, these typically focus either on single type tariff or making strong assumptions regarding In...
Reinforcement Learning has long been employed to solve sequential decision-making problems with minimal input data. However, the classical approach requires a large number of interactions an environment learn suitable policy. This problem is further intensified when multiple autonomous agents are simultaneously learning in same environment. The teacher-student aims at alleviating this by integrating advising procedure process, which experienced agent (human or not) can advise student guide...
The deployment of security services over Wireless Sensor Networks (WSN) and IoT devices brings significant processing energy consumption overheads. These overheads are mainly determined by algorithmic efficiency, quality implementation, operating system. Benchmarks symmetric primitives exist in the literature for WSN platforms but they mostly focused on single or systems. Moreover, not up to date with respect implementations and/or systems versions which had progress. Herein, we provide time...
Although Reinforcement Learning (RL) has been one of the most successful approaches for learning in sequential decision making problems, sample-complexity RL techniques still represents a major challenge practical applications. To combat this challenge, whenever competent policy (e.g., either legacy system or human demonstrator) is available, agent could leverage samples from (advice) to improve sample-efficiency. However, advice normally limited, hence it should ideally be directed states...
The COVID-19 pandemic underscored the promise of monoclonal antibody-based prophylactic and therapeutic drugs
Most previously authorized clinical antibodies against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) have lost neutralizing activity to recent variants due rapid viral evolution. To mitigate such escape, we preemptively enhance AZD3152, an antibody for prophylaxis in immunocompromised individuals. Using deep mutational scanning (DMS) on the SARS-CoV-2 antigen, identify AZD3152 vulnerabilities at antigen positions F456 and D420. Through two iterations of computational design...
Pollinators play a key role in biodiversity conservation, since they provide vital services to both natural ecosystems and agriculture. In particular, bees are excellent pollinators; therefore, their mapping, classification, preservation help promote conservation. However, these tasks difficult time consuming there is lack of classification keys, sampling efforts trained taxonomists. The development tools for automating assisting the identification bee species represents an important...
Reinforcement learning (RL) is a widely known technique to enable autonomous learning. Even though RL methods achieved successes in increasingly large and complex problems, scaling solutions remains challenge. One way simplify (and consequently accelerate) exploit regularities domain, which allows generalization reduction of the space. While object-oriented Markov decision processes (OO-MDPs) provide such opportunities, we argue that process may be further simplified by dividing workload...
Driven by recent developments in the area of Artificial Intelligence research, a promising new technology for building intelligent agents has evolved. The is termed Deep Reinforcement Learning (DRL) and combines classic field (RL) with representational power modern approaches. DRL enables solutions difficult high dimensional tasks, such as Atari game playing, which previously proposed RL methods were inadequate. However, these solution approaches still take long time to learn how actuate...
Autonomous agents are increasingly required to solve complex tasks; hard-coding behaviors has become infeasible. Hence, must learn how tasks via interactions with the environment. In many cases, knowledge reuse will be a core technology keep training times reasonable, and for that, able autonomously consistently from multiple sources, including both their own previous internal other agents. this paper, we provide literature review of methods in Multiagent Reinforcement Learning. We define an...
While reinforcement learning (RL) has helped artificial agents solve challenging tasks, high sample complexity is still a major concern. Inter-agent teaching -- endowing with the ability to respond instructions from others been responsible for many developments towards scaling up RL. RL that can leverage learn tasks significantly faster than cannot take advantage of such instruction. That said, inter-agent paradigm presents new challenges due to, among other factors, differences between...
The optimal design of power converters often requires a huge number simulations and numeric analyses to determine the parameters. This process is time-consuming results in high computational cost. Therefore, this paper proposes deep reinforcement learning (DRL)-based optimization algorithm optimize parameters for using neural network (DNN)-based surrogate model. model can quickly estimate efficiency from input without requiring any simulation. proposed includes two major steps. In first...
Learning to solve sequential decision-making tasks is difficult. Humans take years exploring the environment essentially in a random way until they are able reason, difficult tasks, and colla
Power converters are pervasive in modern electronic component design. They can be found all devices from household appliances and cellphone chargers to vehicles. Currently, designing new circuit topologies is hard because it requires human expertise based on experience difficult automate. However, artificial-intelligence-assisted design significantly facilitate the development of power and/or improve final result. Intelligently designed highly efficient have a significant effect many...
The ongoing industrialization and rising technology adoption around the world are leading to ever higher energy consumption. benefits of electrification enormous, but growing demand also comes with challenges respect associated greenhouse gas emissions. Although continuing progress in research has brought up new technologies generation, storage, distribution, most those focus on increasing efficiency individual components. Work integration coordination abilities between components...
Although Reinforcement Learning methods have successfully been applied to increasingly large problems, scalability remains a central issue. While Object-Oriented Markov Decision Processes (OO-MDP) are used exploit regularities in domain, Multiagent System (MAS) divide workload amongst multiple agents. In this work we propose novel combination of OO-MDP and MAS, called Process (MOO-MDP), so as accrue the benefits both strategies be able better address issues. We present an algorithm solve...
Understanding the degree of humanness antibody sequences is critical to therapeutic development process reduce risk failure modes like immunogenicity or poor manufacturability. We introduce AbBERT, a transformer-based language model trained on up 20 million unpaired heavy/light chain from Observed Antibody Space database. first validate AbBERT using novel “multi-mask” scoring procedure demonstrate high accuracy in predicting complementary determining regions—including challenging...