Google to Open Dopamine Source Code

The tool for neural network training is based on TensorFlow, a library for machine learning
30 August 2018   352

Google Brain Team published the source code of the Dopamine framework, which allows the implementation of training with reinforcement for neural networks. The repository contains 15 Python files with documentation. The tool is based on TensorFlow, a library for machine learning.

The framework is based on the Arcade Learning Environment platform, which evaluates the performance of AI using video games. Developers also got access to sets of source data for training and tests on 60 games supported by the platform. This approach makes it possible to standardize the process of working with neural networks and to obtain reproducible results.

Dopamine supports 4 learning models: deep Q-learning, C51, Implicit Quantile Network and a simplified version of Rainbow.


Simultaneously with the placement of the source code, Google launched a website with tools to visualize the process of interacting with AI via Dopamine. The site supports work with multiple agents simultaneously, provides access to statistics, training models and planning through TensorBoard.

Pablo Samuel Castro and Marc G. Bellemare, Google Brain Team researchers expressed the hope that the flexibility and ease of use of the tool developed by their group will inspire developers to try out new ideas.

This is not Google's first step towards increasing the availability of tools for neural networks. In 2017, the company announced the launch of the project, a project to democratize the achievements in the field of machine learning.

Get more info at GitHub.

AI to be Used to Create 3D Motion Sculptures

The system developed by the MIT and Berkeley scientists is called MoSculp and is based on artificial inteligence
21 September 2018   119

MoSculp, the joint work of MIT scientists and the University of California at Berkeley, is built on the basis of a neural network. The development analyzes the video recording of a moving person and generates what the creators called "interactive visualization of form and time." According to the lead specialist of the project Xiuming Zhang, software will be useful for athletes for detailed analysis of movements.

At the first stage, the system scans the video frame-by-frame and determines the position of key points of the object's body, such as elbows, knees, ankles. For this, scientists decided to resort to the OpenPose library, developed by the Carnegie Mellon University. Based on the received data, the neural network compiles a 3D model of the person in each frame, and calculates the trajectory of the motion, obtaining a "motion sculpture".

At this stage, the image, according to the developers, suffers from a lack of textures and details, so the application integrates the "sculpture" in the original video. To avoid overlapping, MoSculp calculates a depth map for the original object and the 3D model.

MoSculp 3D Model
MoSculp 3D Model

The operator can adjust the image during the processing, select the "sculpture" material, color, lighting, and also what parts of the body will be tracked. The system is able to print the result using a 3D printer.

The team of researchers announced plans to further develop the MoSculp technology. Developers want to achieve from the processing system more than one object on the video, which is currently impossible. The creators of the technology believe that the program will be used to study group dynamics, social disorders and interpersonal interactions.

The principle of creating a 3D model based on human movements has been used before. For example, in August 2018, scientists at the same University of California at Berkeley demonstrated an algorithm that transfers the movements of one person to another.