Dwango Media Village conducts research in the fields of
machine learning, computer vision, computer graphics, audio
processing and game AI.
Research
Hand Generation in Mahjong using Diffusion Process Model
Reproduction Experiment of Video PreTraining (VPT) for Learning to Play Games from Human Play Videos
Generative Probabilistic Image Colorization
Reinforcement Learning System "Marltas" for RPG Atsumaru Posted Games
Reimplementation of the Reinforcement Learning System SPIRAL to Generate Drawing Methods from Images
Team Strategy Analysis Based on a Win-Loss Prediction Model from Splatoon Koshien Streaming Videos
Reinforcement Learning for Artificial Life on mobile devices
Voice Conversion in Two Stages: Transformation and High-Resolution Enhancement
Image Thumbnail Recommendation System Using Tag Information and Comment Density
Article
Alignment Tool pydomino Based on Distinctive Features of Phonemes
Evaluating the Speed Improvement of Deep Learning with TPUs
MarltasCore, a distributed & asynchronous DQN implementation
Animation Generation from a Single Character Front Image
A Voice Conversion System That Can Transform Any Voice into 100 Voices
Attending the International Conference AAAI2020
Sponsored the Marltaus Award at the 2020 Indie Game Festival Newcomer Award
AA-chan, the ASCII Art Detector
A 24-hour live broadcast of Marltas' learning was held
nico-opendata
We are providing a wide variety of data from our services
to
be used for academic purposes.
Recruit
We are hiring machine learning engineers and researchers.
See
the recruit page for more information.