Between “Expectation” and “Anxiety”
The large prologue of Fourth Industrial Revolution always brings up two topics hand in hand- Big Data and Artificial Intelligence. Our reaction to these two major topics would be two contradictory feelings.
Those would be expectation towards the future mankind, along with anxiety due to ambiguity as an individual. This session we would like to open up a discussion with these co-existing sentiments in the age of Big Data and AI.
Ailys seeks to provide values generated by one of AI’s threads, machine learning, to numerous companies over different industries. We could see for ourselves the customers undergoing the two conflicting reactions in actual business fields- that’s precisely how we started our discussion.
“Ailys pursues AI techniques that can achieve collective growth with our users.”
AI questions the user value
Throughout history, advancement in technology has always been aimed towards human survival or convenience. In other words, the most critical subject of discussing technical advancement has been the mankind. However, along with the appearance of AI has become ambiguous the real line that tells apart humans and non-humans. That was the cue that reeled us to focusing on the technology’s excellence rather than inquiring who the users would be or how they would operate. In that case, could AI be made immaculate even when its users lacked skills or knowledge?
AI will continue to replace humans in many areas, especially considering that it can already recognize images and comprehend inter-human conversations. Nonetheless, Ailys strongly believes in the strength underlying in the intrinsic human value, so we developed an AI solution for those working in current business fields. Why so? We traced the answer from the essence of AI, attaining particular “intelligence” by training on given data. In that sense, what characteristics should that data hold? Would merely holding vast amount of data qualify for a “good training set”?
Of course, the answer is no. It’s easier to understand once there’s an example related to us humans. Direction of training is extremely important in knowledge acquisition. Assuming someone assigned us a task called ‘A’, blindly scarfing down through the existing sources like books and reports is not the optimal way to solve it. Someone wise enough would rather train on tasks that are similar to Task ‘A’. The same approach should be applied to AI. Even if machines and technology could process much more elaborate judgements compared to humans, they still require human guidelines to narrow down the type of data in order to solve the problems.
This is precisely where we seek for user value. User value is enormously impactful in data analyzing process, as it’s based on machine learning that operates on data table information. It’s on current business expert’s hands to determine the accuracy of the machine learning model’s predictions- which training data should be selected, and whether to execute an experiment to improve the model engine or not.
Freedom from the technology, DAVinCI LABS
Predictive modeling is not a domain in which you can achieve the best results from a mere one-time operation. What’s important is the “performance advancement through repetitive training”, learning diverse models via various algorithms, and then choosing one with the highest performance. In this case, who should be the one in charge of such “repetitive training”? Would it be enough to be prepared with groups of data experts? As you could probably assume from the information above, the answer is surely “no”. Like the theory of an architect being skilled with architecture, experts who are experienced with repeated experiments on the particular problem would be the ones guaranteeing the best output.
Consequently, at this point we, Ailys, came up with our AutoML solution “DAVinCI LABS”. Inspired by how countless experiments conducted inside Davinci’s laboratory benefitted the advancement of the mankind, we questioned how much potential a predictive model would carry when driven by machine learning techniques, composed by experts in the business. Our story started from a somewhat generic idea and was solidified into a productive outcome.
“DAVinCI LABS” is an automatic machine learning solution containing all experiences and know-hows of data scientists. We dream our future to consist of an environment where all the following are possible: Providing a means of approach towards “machine learning” technologies, seemingly distant for those who aren’t accustomed to data science, for all experts over different domains. Also, advancing the predictive model so as to optimize it for respective domains; not to mention opening up a door of new opportunities for business experts by freeing them from technological requirements, consequently securing an environment where AI and current business experts could grow together.
When we discuss artificial intelligence, we commonly come across the expression “Data-driven”- it’s an expression emphasizing that AI works towards highly advanced training and accurate prediction based on data, not just on a rough guess. Additionally, we would like to underline another value, “Human-driven AI”. Although AI is the subject that trains on data, it eventually boils down to humans for selecting and providing a good-quality dataset. Along with Ailys striving on the collaborative standpoint of AI and humans, we hope you could head to a path of success on your career and business with “DAVinCI LABS”.
