Training machine learning algorithms without the appropriate tools and technical know-how can be downright frustrating.
Fortunately, for those without the necessary tools to execute this task successfully, Baidu is offering an online tool dubbed as EZDL that is currently in beta stage. The recently unveiled tool is designed to make it easy for almost anyone to create, design and deploy artificial intelligence models without having to write any code.
According to the general manager of Baidu’s artificial intelligence division, Youping Yu, the EZDL tool was created with various qualities in mind including security, ease of use and performance. He also said that it targets three main categories of machine learning including sound classification, object detection, and image classification.
Yu added that EZDL is aimed at small and medium-sized enterprises, specifically those with the objective of “breaking down the barrier” in a bid to enable everyone to access artificial intelligence(AI) in the most convenient and equitable way.“
EZDL needs about 20-10 images or even over 50 audio files that are assigned to a given label in a bid to train a model. Training time takes about 15-60 minutes. Furthermore, Baidu claims that over two-thirds of models reach more than 90% accuracy scores.
Generated algorithms cannot only be deployed in the cloud but also accessed through an API. Alternatively, such algorithms can be downloaded as a software development kit that supports Android and iOS among other operating systems.
According to Yu, Idcool, a home decorating website, utilized EZDL in training systems that automatically recognize the style and design of a room with 90% accuracy. Also, an undisclosed medical institution used it to create a detection model for conducting blood test microscope imagery.
Furthermore, a security monitoring company leveraged it in making a sound-detecting algorithm that can identify “abnormal” audio patterns that may represent a break-in.
“As one of the global leaders in artificial intelligence (AI), Baidu embodies its mission of making a complex world simpler through technology,” said Yu.
He went on to say, “Baidu has continued to lead and promote the development and applications of AI. Through our Global Business Unit, we are devoted to sharing Baidu’s core technologies, capabilities, and resources with the global community…We are excited to see innovative use cases from the global community.
It is our hope that by opening up the platform and democratizing AI capabilities, we will help developers and businesses around the world achieve greater success.”
Baidu has made its artificial intelligence(AI) ambitions clear in the past two years since launching Baidu Brain, its enterprise AI’s eponymous platform. Currently, the company claims that over 600, 000 developers are using Brain 3.0, which is the newest version that was launched in July 2018 for 110 AI services across 20 different industries.
Back in July, Baidu announced Kunlun, an AI chip made to handle models for both datacenter processing and on-device edge computing.
Once it launches in the approaching months, the company says that it will be in a position to attain 260 tera-operations per second and 512 GB per second memory bandwidth. Also, in August, Baidu revealed that its conversational AI assistant dubbed DuerOS had hit an install base of 100 million devices, an increment from 50 million.
Source VentureBeat
