In the second artificial intelligence event in the industry, it was raised; Assessing the transformative effects of technology, from medicine and education to industry and marketing

This event was held for the second year in a row with the aim of creating a space for the exchange of ideas of experts and activists in the field of artificial intelligence, machine learning and data science in the country and benefiting from the experiences that the speakers of the conference have gained in facing the challenges of developing artificial intelligence and data science products. It was done to both help networking in this nascent field and facilitate the development process of this new technology.

In this one-day event, speakers from Google, Meta, Microsoft, IBM and Mofid Broker were present and after the end of each presentation, they answered the questions of the attendees.

Also, in the middle of this event, a discussion panel was held on the topic of “the impact of ChatGPT and related technologies on the future of businesses”, where the experience and challenges of some speakers were conveyed to the audience in the form of a dialogue.

In the second artificial intelligence experience-oriented event in industry (AIX), Sohail Moghadam, Vice President of Digital Kala Technology, Hamidreza Vaezi, Senior Data Scientist of Meta, Mohammad Shakohi Yekta, Senior Data Scientist of Microsoft, Arzoo Keshavarez, former technical director of Google, Kianoosh Mokhtarian, former technical director of Google, Mehdi Salmani Head of IBM Watson data team, Ali Zarezadeh, head of data and artificial intelligence team of Mofid Agency, Zainab Barzegar, member of Iran University of Medical Sciences, Mohammad Hossein Rehban, member of Sharif University of Technology, were present as speakers.

In the first presentation of this event, Kianoush Mokhtarian, the former technical director of Google, gave explanations about display ads in the web environment and the models that are designed for their strategy. Referring to the machine learning models in the operational environment, he said: the problem in these models is deployment in the operational environment. A model cuts across architecture and a model cuts across traffic, and ultimately all of these models ultimately help your display ads perform best for you on the web.

As the second speaker, Sohail Moghadam, vice president of technology at Digikala, spoke about his experiences in using artificial intelligence at Digikala. Pointing out that the consumption of boxes in Digikala is very high, he added: Until a few years ago, I kept thinking to myself, is it necessary to use so many natural resources and paper? This thought made us think of a way to design the box according to the product and not necessarily use the box at all. For this, we took help from artificial intelligence. Artificial intelligence helps us to produce the boxes according to the product and avoid paper waste.

He also continued: With the help of artificial intelligence, the schedule of sending goods has become more regular and logical, and this modification in the long schedule also pollutes the environment less.

According to Moghadam, artificial intelligence is used in Digikala to check the quality or authenticity of the goods sold by a seller in Digikala, or the comments made by the customer and the photos and videos he sends. Moghadam stressed: Artificial intelligence as an open source tool is a kind of easy and restrained that can make the affairs of a business easy.

“Artificial intelligence in the operating room” was the topic of the speech of Zainab Barzegar, a faculty member of the Iran University of Medical Sciences, at the second conference of artificial intelligence in industry. Referring to the history of surgery, this computer engineer said: Until 150 years ago, anesthesia did not exist. The person who was going to be operated on had to endure the pain by injecting narcotics. Or more interestingly, there was no such thing as a surgeon until the 19th century. Even in the 19th century, when surgery began to take shape, there was no culture of disinfection, and the doctor would suddenly start surgery without being equipped with special facilities, and these conditions caused a small number of people to come out of the operating room alive and well.

The faculty member of Iran University of Medical Sciences further added: Little by little, with the discovery of disinfectants, X-rays and ultrasonic rays, the surgical process progressed in a way and a unique transformation was created in this field. Now we have reached a point where artificial intelligence is constantly present in operating rooms. Machine learning, analysis of images, linguistic understanding and artificial neural network are the most important fields in which artificial intelligence has entered and helps the development of medical science.

Barzegar also mentioned the devices that can completely find the problem in the patient’s body by using artificial intelligence. He added: For example, before any incision is made in the patient’s body, a number of landmarks are created in the patient’s body, which can be matched with the images taken from the patient’s body. With the help of anatomy knowledge and artificial intelligence, the technician can start working more accurately and choose the best point for surgery.

“The evolution of computer vision in the industry in the last decade” was another topic that was discussed at the second artificial intelligence experience-oriented conference in the industry. Hamidreza Vaezi, Chief Data Scientist at Meta Company, started his talk with the focus that it is never too late to enter the field of computer vision.

Regarding the changes in computer vision in the industry in the last decade, he said: At the beginning, we must explain about computer vision, which is an important part of artificial intelligence. Computer vision does three main tasks: first, we need to see what the photo is; Then let’s see where the image is; Then we have to see where the objects we are looking for are in the image.

Vaezi explained about the increase in interest in machine learning over time: In 2013, only a hundred people were working on machine learning at Apple. The year after that, what happened was that we went from 100 people to 400 people, and in 2015, this number for us exceeded 1000 people.

He said: AI models should be balanced in the field of race, gender, etc. Artificial intelligence has been hotly debated in the past years and is set to become more precise in the coming years. Our duty is to advance the technology and pave the way for it. Artificial intelligence can change the game for you in the future. It’s not too late to start computer vision; We have to keep ourselves updated because two years from now will be different from today. We’ve only been in this for ten years, and we’re definitely going to see some interesting things happen in the future in the field of computer vision.

Ali Zarezadeh, the leader of Mofid Brokerage’s data and artificial intelligence team, explained his personal experiences in using ChatGPT: The first days when ChatGPT came and I was using it, it was not really strange to me. The only thing that attracted me to using this language model was the ability to chat with it, because previous models only completed sentences, but ChatGPT had a conversation with the user and its information was gradually completed with the help of a human.

He continued: I also thought early on that ChatGPT is the image version of Google. But little by little I saw that he understands reasoning and solves riddles. I was asked how a language model can reason? I believe that this argument also works on data and designs a function that has a much stronger chance of producing a sentence with a correct argument.

Zarezadeh also pointed to the design of a useful customer support chatbot, with the help of the statistics in the data bank of this collection, and said: The demo version of this chatbot was presented at the 1402 financial industry exhibition, and its final model will be unveiled by the end of summer. According to him, with the help of this technology, up to 50% of the questions of brokerage clients are answered quickly and easily without spending extra time.

In the second half of the experience-oriented event of artificial intelligence in the industry, a specialized panel was held on the topic of the impact of ChatGPT and related technologies on the future of businesses. In this discussion panel, Mohammad Hossein Rehban, a member of the scientific faculty of Sharif University of Technology, Mohammad Shakohi Yekta, a senior data scientist at Microsoft, Sohail Moghadam, vice president of digital technology, Zainab Barzegar, a member of the scientific faculty of Iran University of Medical Sciences, and Ali Zarezadeh, the leader of the data and artificial intelligence team of Mofid Brokerage, were present.

As the first question, Shokohi Yekta asked the attendees at the meeting: “What are the most important uses of productive artificial intelligence in your expertise and job?”

Zainab Barzegar explained about this: producing images is the most important help that productive industrial intelligence gives us. I think that in industry and academia in the not too distant future, AI may even change the course of medical diagnosis.

Pointing out that we use artificial intelligence to help sellers, Sohail Moghadam said: One of the applications of artificial intelligence for DigiKala is to produce a box of the same size as the product to reduce paper consumption and manage costs for the business.

Rehban also believed that artificial intelligence can be used in various ways to write articles.

Another question that was raised in this panel was “Is it true that they say that with the development of artificial intelligence, the future of humanity is in danger?”

In response to this question, Zainab Barzegar said: We should not depend on artificial intelligence and abandon research and reading. We used to spend months reviewing an article. But now the story has changed and this issue is especially bad in the academic environment and I am worried about illiterate graduates.

The monk also believed: In my opinion, if we use the methods in the right way, it is very good, but it is a double-edged sword.

But in response to this question, Moghadam explained: The world is changing and usually problems come along with an industry, and next to these problems, another industry is created that answers these problems. Regarding studying and learning, based on my reviews, learning has taken the form of AI learning and will remain so. It means that teachers should adapt themselves to the way children learn. Learning is not lost. Only its type changes. We have to go for new and new methods. Challenges shake up systems that are in disarray. Today, we are witnessing a fundamental transformation that threatens the future stability of humanity (at first), but eventually we get used to it and find a way to live better with it.

Ali Zarezadeh went on to say: I think the work that ChatGPT is doing today is good, and we, as human resources, should go for higher level work. Just like all the revolutions we went through. Is it bad? Now we don’t plow by ourselves and the machine does it. Artificial intelligence is exactly that. He added: A software engineer must have some basic skills and advanced skills. In my opinion, we should always know the basics. But don’t waste your time on low-level tasks and leave them to machines.

Arezoo Keshavarz, the former technical director of Google, who was present online at the second artificial intelligence-based experience event in the industry, explained the definition of data analysis science and its relationship with artificial intelligence and its advantages for businesses and various sciences. Keshavarz said: Data analysis is a decision-making tool for businesses and different people. From a very large company that has different employees and departments to a therapist who needs new psychotherapy methods.

He also pointed to the new functions of data analysis and said: Today, we can use the science of data analysis to improve the performance of a search engine.

He also explained: Another application of data analysis science is helping to repair and maintain expensive devices in various industries.

“Unleash the power of data using graph technology and large linguistic models.” This was the title of the last lecture (online) of the second experience-oriented artificial intelligence conference in the industry. JPMorgan’s artificial intelligence director explained to the audience about the benefits of graphs in artificial intelligence. He explained: There are companies and investors in this field and data is obtained from different sources. Suppose you got the data; Now the challenge is how to connect the data together. If you do not have a proper understanding of this market, it is difficult to use the data correctly. Graph helps you to solve these problems.

As the last speaker, he continued to emphasize: In the usual case, you have to do many things to answer a simple question; But when you have a graph, you say I’m looking for this kind of relationship, and in one line you can get information that you could hardly get in the past in a few minutes. The graph also has a flexible structure.