Iris Technology has released a new no-code solution that enables developers and enterprises to train and deploy AI models faster. Much less data and computing power. The platform, webAI, accelerates the AI and computer vision process while giving companies control over their intellectual property.
look: Recruitment Kit: Computer Vision Engineer (TechRepublic Premium)
Starting the first week of January, webAI will be available in a limited beta release. The company ensures its new technology disrupts traditional approaches to AI.
TechRepublic spoke with Iris Technology’s two co-CEOs, James Meeks and David Stout, for a behind-the-scenes look at the company’s new platform, the potential of no-code AI, and its challenges.
Jump to:
webAI: what you can do
Iris Technology has been developing webAI in stealth mode for the past three years. The release of the WebAI platform allows developers and businesses to build models and prototypes quickly and for free before investing in an enterprise license. The new solution significantly reduces deployment time.
look: How low-code and no-code solutions help great developers become even better developers (Tech Republic)
“The biggest benefit is making AI much more accessible and cost-effective,” Meeks said. “Currently, there are only about 300,000 AI experts in the world, so building a team of people who can build AI models is a big challenge. This means teams of developers can build, train, and deploy models with state-of-the-art performance without requiring a deep AI background.”
According to the company, compared to YoloV7, currently considered the fastest and most accurate real-time object detection model for computer vision tasks, webAI requires 1/5 the data to train and 1/3 the training time. am.
Additionally, since training is always free, WebAI enables iterative development, getting models into the field faster and significantly reducing risk. Only about 10% of traditional computer vision AI models have been deployed, and iterations require rebuilding the entire model.
“Most AI platforms today are built on the assumption that big data is the answer to the world’s problems,” Stout said. “webAI throws that assumption out the window. Our radically different approach allows virtually any developer, regardless of budget or previous experience with AI, to create AI models quickly and cost-effectively.” We envision a world where we can train, deploy, and iterate on.”
Key Features of WebAI Beta Release
Key features of the WebAI beta release include:
- Agility and Speed: Rapid curation and deployment with less model training.
- Sensor independent capacity: The trained iris model works regardless of camera type and computer.
- Edge Ready: WebAI models can run on most consumer-grade laptops and do not require cloud computing, so the platform has low computational requirements.
- Data privacy and IP protection: Delivery via blockchain allows customers to build models in their own environment for enhanced security and privacy. Customer data and intellectual property belong to the customer, not to Iris Technology.
- No-code mode and full-code mode: The platform offers no-code and full-code modes for improved accessibility while giving experienced developers full control.
webAI believes that no-code opens the door for AI to solve real-world problems and create disruptive value where traditional AI has been prohibitively expensive and ineffective.
look: State-of-the-art computing platform (Tech Republic)
“Companies are investing billions of dollars in artificial intelligence expertise, computing infrastructure, and data acquisition curation to fuel traditional AI experiments that have about a 13% chance of being deployed.” Meeks explains.
Developers and businesses can develop AI apps “without spending hundreds of thousands of dollars on computing infrastructure, data collection, and curation.”
The No-Code AI Market and Its Importance
webAI addresses the challenges of AI computer vision and the processes required to develop new AI applications. From managing data quality, to choosing app features and training, to deploying and maintaining a solution, developing a new AI app takes time. Many processes are still artisanal and completed manually by data teams.
look: best data quality tools (Tech Republic)
But new AI automation tools, such as advanced feature engineering, are becoming increasingly available for developers to help data professionals streamline production. In this environment, no-code AI is seen as the ultimate automated approach to AI development.
Future Market Insights forecasts that the global no-code AI platform market will reach $38.5 billion in 2032, growing at a CAGR of 28.1%. The market value in 2021 was just $2.58 billion.
The urgent need for automation, the adoption of ML and AI across industries and sectors, the time-consuming and costly factors of building AI from scratch, and the shortage of skilled AI-savvy workers have pushed no-code AI is expected to: keep growing.
look: Adoption Kit: Autonomous Systems Engineer (TechRepublic Premium)
Popular no-code applications include Knack, Bubble, Lansa, RunwayML, and Substack. Tech giants such as Google and Microsoft are also developing no-code AI to enrich their cloud services and attract new customers.
However, despite the potential of new technologies, no-code AI also presents many challenges.
Addressing No-Code AI Challenges
No-code AI shares some commonalities with traditional AI in terms of performance. For example, model drift, where AI applications produce inefficient or inaccurate results due to changes in environmental data, can affect both types of technologies. However, the no-code AI industry must also overcome other negative perceptions associated with products such as black box AI.
black box AI
Black-box AI, often associated with no-code AI, is when an AI application produces advanced results, but the inner mechanisms of how the algorithm achieved the results are unknown. Black-box models have been criticized for their lack of transparency and inability to verify results.
With this concern in mind, TechRepublic asked Iris Technology how webAI addresses the challenges of black box AI and provides transparency.
“Users with more expertise can work in a full-code environment where they can build elements and workflows from scratch,” Strout explains. “While webAI’s new architecture, Deep Detection, is not open source, the platform is incredibly accessible and transparent. and iterate over them, as well as own and control their models and all their inputs and outputs.”
synthetic data
Synthetic data is another no-code AI and ML trend that is gaining momentum. It is increasingly used for algorithms that require biometric, video, and photo data.
Data for AI projects are difficult to obtain because they must be obtained with consent from the creator or owner and must be diverse to avoid biased and discriminatory results. , and can be generated in large numbers to train AI apps.
look: Machine Learning and Artificial Intelligence Certification Bundle (TechRepublic Academy)
However, developers question whether the synthetic data match the quality of the real information. I also question the ability to create different databases and functions.
“While webAI does not currently utilize synthetic data, we believe there are use cases where synthetic data is a good option,” Stout said. “webAI has the essentials for training AI models. If you are using one of our proprietary architectures, there are several scaling benefits in parallel to enrich your dataset. will occur.”
Data preparation and model drift
Preparing data for ML and AI is also a hot topic, as data must meet the highest standards for algorithms to work effectively. Inconsistent, outdated, or omitted data can cause model collapse and drift.
Stout believes webAI is very transparent when it comes to data quality standards.
“When we talk about high-quality data in WebAI, we often refer to well-defined, clean data,” says Stout. “For most applications, sensors are not gates. Sensors are typically under-informed and mislabeled can prevent model deployment from reaching its potential, but most raw cameras You can get his feed without pre-cleansing.”
look: Data Cleansing: Cheat Sheet (Tech Republic)
Monitoring AI applications is essential for enterprises. This is especially important in modern enterprises, where unforeseen events, market and supply chain disruptions, and environmental issues can lead to significant changes in data.
According to Stout, Iris Technology built webAI as an AI tool to provide explainability to creators.
“Once a model is deployed in a workflow, users can monitor it within the IDE,” Stout explains. “For example, a product deployed using webAI can be reviewed in real time by the developer or team utilizing the model.”
To monitor your application, the interface provides real-world feedback and metrics to ensure optimal model performance throughout its lifecycle.
The future of no-code AI in the workplace
No-code AI will undoubtedly allow many companies to reduce costs and take advantage of cutting-edge technology while deploying already-tested algorithms, but no-code AI requires data teams and advanced skills. Will it replace the workers who have?
look: Artificial Intelligence Ethics Policy (TechRepublic Premium)
In the opinion of the Iris Technology team, no-code AI benefits data professionals and non-data professionals alike.
“No-code AI will give more people the ability to train, deploy, and iterate models. And webAI’s novel approach will enable data scientists and engineers to do it faster and more cost-effectively. It means you can do it with ,” says Meeks. “Far from replacing human input, we believe there will be a growing demand for human expertise and creativity to bring AI into new areas.”
Read the following: 8 Most Innovative AI and Machine Learning Companies (Tech Republic)
