Sotheby’s sold the original World Wide Web source code for $5.4 million — © AFP Martin BUREAU
AI without code offers a conversational interface based on artificial intelligence and machine learning to perform data analysis. No need to write query code, combined with an intuitive interface, allows non-technical professionals to access data to explore, query, analyze, test hypotheses, and create predictions based on their data.
This flexibility reduces the need to rely on in a team of data scientists to provide insights that enable companies to make smoother data-driven business decisions. To get an idea of the development of technology, digital magazine spoke with Kevin Chen, Chief Data Scientist Experian DataLabs in North America.
Digital Journal: What is a Codeless AI (Artificial Intelligence) System? And why is it important?
Kevin Chen: Codeless systems provide a user interface that allows users to create application functionality without writing programming code. Conceptually, this is similar to what you see is what you get (WYSIWYG) HTML website design editor, offering a visualization layer on top of the code that makes it much easier to manipulate that code. Codeless AI takes the same approach to data science and AI/machine learning, except in a much more complex way.
Codeless AI, one of the cutting-edge technologies of our R&D division Experian DataLabs, offers a conversational AI-based interface based on machine learning to perform data analysis. No need to write query code, combined with an intuitive interface, allows non-technical professionals to access data to explore, query, analyze, test hypotheses, and create predictions based on their data. This flexibility reduces the need to rely on a team of data scientists to get insights. The idea is to make the data available to product managers, sales teams, store owners, and anyone else who might benefit from it.
DJ: AI without code eliminates the need for data scientists?
Chen: No, because codeless artificial intelligence still requires pre-coding to provide a hidden view of the data and intelligence. For example, he needs ML data scientists in natural processing/understanding/generation areas to develop the ability to understand user intent for any requested analysis. In addition, code-free artificial intelligence benefits from algorithms developed by data scientists to automatically detect trends and patterns or perform root cause analysis.
In short, AI without code is only as good as the level of intelligence that is built into it (i.e., coded). Even with practical business challenges, when an organization has a significant amount of domain knowledge, writing back-end code for a non-AI-coded UI can be a challenge due to the nuances of the data, how the data is to be used, or some other unique application. within the organization’s operations.
Until there is real AI that does its own analysis, users will still need a basic understanding of what AI can and cannot do without code. Asking him to predict customer churn would seem simple, except that the definition of churn may vary depending on the business application. For example, a bank customer actually or literally stops using his bank account, but it remains open. In practice, such a scenario should be treated as an exhaustion, but it is often not marked as closed in the bank’s system. The presence of artificial intelligence without code simply uses the closed account flag in the data, since the intended target definition for predicting depletion will in this case lead to a less desirable result for the bank.
As code-free AI enables non-technical users to gain valuable insights from their data for data-driven decision making, data scientists must remain a key partner on this journey, helping to validate any data assumptions made and guide analysis. At the same time, the efficiency provided by AI without code will free data scientists from routine data analysis and routine reporting, so their valuable time can be devoted to solving complex and important business problems.
DJ: What is the right way to set up an artificial intelligence system without code to enable such open use?
Chen: Several fundamental functional building blocks are required, starting with the ability to perform common ML/AI tasks such as data ingestion, data cleaning, data quality control, feature extraction, model training using various ML techniques, parameter search, and model evaluation. .
Then there should be a business logic layer that allows codeless artificial intelligence to help users solve certain types of business problems, as well as perform simple general machine learning tasks to make predictions based on a set of data points, identify contributing factors. to a business opportunity or challenge, etc.
The system then needs to have an intelligent presentation layer that automatically creates the most appropriate data visualization to provide users with the requested information.
To combine these AI capabilities with voice activation for queries, codeless AI needs powerful natural language processing/understanding/generation (NLP/NLU/NLG) capabilities to interact with the user using plain English or another desired spoken language. This function is needed to understand the user’s question and invoke the correct parsing and/or even directly convert the request into executable code.
This functionality goes way beyond Alexa and Siri-level voice recognition and activation, or even existing task-oriented conversational AI/chatbot solutions that can look up weather conditions or book flights, for example. Deep learning based solutions are needed to try to understand what users want and then match that user’s intent with the data available. The challenge is to match the domain knowledge of an artificial intelligence system without code about the data at its disposal with the knowledge of the user making the request. If the user believes that the system should have the appropriate data and present it, but the query returns insufficient results, a disconnect will occur.
DJ: Please give an example of how artificial intelligence can work without code.
Chen: Experian DataLabs is constantly on the lookout for disruptive forces in the market that can benefit from the confluence of raw data and machine learning, such as offering greater access to advanced data analysis, research and development for non-technical users.
To facilitate this type of access and allow users to experiment with data queries, Experian DataLabs has developed the Ascend analytic sandbox. It is an advanced analytics tool that uses nearly two decades of anonymous credit data from over 220 million consumers, as well as business, property and other alternative data sources.
This deep, extensive repository of anonymous consumer data allows data scientists and non-technical decision makers to study behavior, trends, demographics and more. Experian’s Ascend Interact is a no-code AI conversational analytics agent that Experian DataLabs is developing to help users query and perform data analysis in this sandbox using plain English.
Ascend Interact leverages NLU and NLP deep learning to give business decision makers the ability to directly interact with massive amounts of data and potentially combine it with their organization’s data without having to pass it through a team of data scientists.
From a business intelligence perspective, the product can provide dynamic charts and data-driven analysis based on what the user needs at the moment.
DJ: What does artificial intelligence do next without code?
Chen: The promise of codeless AI and its applications is significant and will continue to provide data analytics and insights to business decision makers. The key will be to integrate the technology into more and more business applications across industries. Early adopters are likely to come from areas where there is a lack of data scientists/analysts, and the domain knowledge required for analysis can be easily implemented into artificial intelligence without code. As the technology advances and a new wave of non-technical professionals begin to understand its potential to drive business growth and mitigate potential risks, codeless AI adoption will flourish.