A5011169398- Reinforcement Learning in Robotics
- Robot Manipulation and Learning
- Music and Audio Processing
- Speech and Audio Processing
- Machine Learning and Data Classification
- Transportation and Mobility Innovations
- Muscle activation and electromyography studies
- Adversarial Robustness in Machine Learning
- Prosthetics and Rehabilitation Robotics
- Music Technology and Sound Studies
- Explainable Artificial Intelligence (XAI)
- Smart Agriculture and AI
- Plant Disease Management Techniques
- Plant Virus Research Studies
- Vehicle Routing Optimization Methods
- Railway Systems and Energy Efficiency
- Zebrafish Biomedical Research Applications
- Neuroendocrine regulation and behavior
- Natural Language Processing Techniques
- Zoonotic diseases and public health
- Receptor Mechanisms and Signaling
- Virology and Viral Diseases
- Energy Load and Power Forecasting
- Forecasting Techniques and Applications
- Culinary Culture and Tourism
Korea University
2023-2024
Centro Universitário de João Pessoa
2023-2024
Institute for Design Problems in Microelectronics
2023-2024
RIKEN
2023-2024
Central Conservatory of Music
2023-2024
École Polytechnique Fédérale de Lausanne
2016-2021
International Institute of Information Technology, Hyderabad
2017
European Organization for Nuclear Research
2016
Crop diseases are a major threat to food security, but their rapid identification remains difficult in many parts of the world due lack necessary infrastructure. The combination increasing global smartphone penetration and recent advances computer vision made possible by deep learning has paved way for smartphone-assisted disease diagnosis. Using public dataset 54,306 images diseased healthy plant leaves collected under controlled conditions, we train convolutional neural network identify 14...
Translating satellite imagery into maps requires intensive effort and time, especially leading to inaccurate of the affected regions during disaster conflict. The combination availability recent datasets advances in computer vision made through deep learning paved way toward automated image translation. To facilitate research this direction, we introduce Satellite Imagery Competition using a modified SpaceNet dataset. Participants had come up with different segmentation models detect...
Vaccine refusal can lead to renewed outbreaks of previously eliminated diseases and even delay global eradication. Vaccinating decisions exemplify a complex, coupled system where vaccinating behavior disease dynamics influence one another. Such systems often exhibit critical phenomena-special close tipping point leading new dynamical regime. For instance, slowing down (declining rate recovery from small perturbations) may emerge as is approached. Here, we collected geocoded tweets about...
Abstract Clinical research in autism has recently witnessed promising digital phenotyping results, mainly focused on single feature extraction, such as gaze, head turn name-calling or visual tracking of the moving object. The main drawback these studies is focus relatively isolated behaviors elicited by largely controlled prompts. We recognize that while diagnosis process understands indexing specific behaviors, ASD also comes with broad impairments often transcend behavioral acts. For...
The automatic recognition of food on images has numerous interesting applications, including nutritional tracking in medical cohorts. problem received significant research attention, but an ongoing public benchmark non-biased (i.e., not scraped from web) data to develop open and reproducible algorithms been missing. Here, we report the setup such a using publicly available sourced through mobile MyFoodRepo app used Through four rounds, released MyFoodRepo-273 dataset constituting 24,119...
This paper summarizes the music demixing (MDX) track of Sound Demixing Challenge (SDX'23).We provide a summary challenge setup and introduce task robust source separation (MSS), i.e., training MSS models in presence errors data.We propose formalization that can occur design dataset for systems two new datasets simulate such errors: SDXDB23_LabelNoise SDXDB23_Bleeding 1 .We describe methods achieved highest scores competition.Moreover, we present direct comparison with previous edition (the...
Learning in multi-agent scenarios is a fruitful research direction, but current approaches still show scalability problems multiple games with general reward settings and different opponent types. The Multi-Agent Reinforcement MalmÖ (MARLÖ) competition new challenge that proposes this domain using 3D games. goal of contest to foster agents can learn across types, proposing as milestone the direction Artificial General Intelligence.
Though deep reinforcement learning has led to breakthroughs in many difficult domains, these successes have required an ever-increasing number of samples. As state-of-the-art (RL) systems require exponentially increasing samples, their development is restricted a continually shrinking segment the AI community. Likewise, cannot be applied real-world problems, where environment samples are expensive. Resolution limitations requires new, sample-efficient methods. To facilitate research this...
This paper summarizes the cinematic demixing (CDX) track of Sound Demixing Challenge 2023 (SDX'23). We provide a comprehensive summary challenge setup, detailing structure competition and datasets used. Especially, we detail CDXDB23, new hidden dataset constructed from real movies that was used to rank submissions. The also offers insights into most successful approaches employed by participants. Compared cocktail-fork baseline, best-performing system trained exclusively on simulated Divide...
Efficient automated scheduling of trains remains a major challenge for modern railway systems. The underlying vehicle rescheduling problem (VRSP) has been focus Operations Research (OR) since decades. Traditional approaches use complex simulators to study VRSP, where experimenting with broad range novel ideas is time consuming and huge computational overhead. In this paper, we introduce two-dimensional simplified grid environment called "Flatland" that allows faster experimentation. Flatland...
We trained a computer vision algorithm to identify 45 species of snakes from photos and compared its performance that humans. Both human is substantially better than randomly guessing (null probability correctly given classes = 2.2%). Some (e.g., Boa constrictor ) are routinely identified with ease by both humans, whereas other groups uniform green snakes, blotched brown snakes) confused. A complex largely molecular delimitation (North American ratsnakes) was the most challenging for vision....
Background: Wet markets are selling fresh meat and produce. critical for food security sustainable development in their respective regions. Due to cultural significance, they attract numerous visitors consequently generate tourist-geared information on the Web (ie, social networks such as TripAdvisor). These data can be used create a novel, international wet market inventory support epidemiological surveillance control settings, which often associated with negative health outcomes....
Multi-agent behavior modeling aims to understand the interactions that occur between agents. We present a multi-agent dataset from behavioral neuroscience, Caltech Mouse Social Interactions (CalMS21) Dataset. Our consists of trajectory data social interactions, recorded videos freely behaving mice in standard resident-intruder assay. To help accelerate studies, CalMS21 provides benchmarks evaluate performance automated classification methods three settings: (1) for training on large datasets...
We here summarize our experience running a challenge with open data for musical genre recognition. Those notes motivate the task and design, show some statistics about submissions, present results.
In this paper, we propose a novel methodology and design to contribute towards the achievement of 17 Sustainable Development Goals (SDGs) adopted by member states United Nations for better more sustainable future all. We particularly focus on achieving SDG 4.7—using education ensure all learners acquire knowledge skills needed promote development. describe crowdsourced approach monitor issues at local level, then use insights gained indicate how learning can be achieved entire community....
The Flatland competition aimed at finding novel approaches to solve the vehicle re-scheduling problem (VRSP). VRSP is concerned with scheduling trips in traffic networks and of vehicles when disruptions occur, for example breakdown a vehicle. While solving various settings has been an active area operations research (OR) decades, ever-growing complexity modern railway makes dynamic real-time virtually impossible. Recently, multi-agent reinforcement learning (MARL) successfully tackled...
To facilitate research in the direction of sample efficient reinforcement learning, we held MineRL Competition on Sample Efficient Reinforcement Learning Using Human Priors at Thirty-third Conference Neural Information Processing Systems (NeurIPS 2019). The primary goal this competition was to promote development algorithms that use human demonstrations alongside learning reduce number samples needed solve complex, hierarchical, and sparse environments. We describe competition, outlining...