Powerful Examples of How Artificial Intelligence is Improving Our World w/ Dr. Djamila Amimer (Episode 19) #DataTalk

Listen to the podcast:

Every week, we talk about important data and analytics topics with data science leaders from around the world on Facebook Live.  You can subscribe to the DataTalk podcast on iTunesGoogle PlayStitcherSoundCloud and Spotify.

In this week’s #DataTalk, we’re talking with Dr. Djamila Amimer about ways artificial intelligence is used for good across different industries. Dr. Amimer is the CEO and Founder of Mind Senses Global.

This data science video series is part of Experian’s effort to help people understand how data-powered decisions can help organizations develop innovative solutions.

Some AI companies mentioned in this episode:

    • Education US: Knewton, an adaptive learning platform, which leverages machine learning to personalize content to improve academic outcomes.
    • Agriculture: San Francisco based company Plenty It uses IoT sensors and machine learning to grow crops vertically indoors using only light, water and nutrients.
    • Forest monitoring project: use machine learning to identify factors that contribute to forest losses in the Congo and the Amazon
    • Suicide prevention: New York City based nonprofit Crisis Text Line analysed millions text messages to determine the words that are most associated with a high risk of suicide. Crisis Text Line is now able to respond to 94 percent of high-risk texters in less than five minutes.
    • Other examples: include the using of machine learning and AI to predict natural disaster such as earthquakes. AI help achieve the UN sustainable goals (17) X prize sponsored by IBM Watson.
    • AI and cancer detection figures:
      Skin cancer melanomas: machine find 95% of melanomas vs. 86% dermatologists
      o Breast cancer: researchers at Imperial College London are working with DeepMind to improve the accuracy of breast cancer screening.
      o Colorectal cancer: detects early stage with 86% accuracy

To keep up with upcoming events, join our Data Science Community on Facebook or check out the archive of recent data science videos. To suggest future data science topics or guests, please contact Mike Delgado.

Mike Delgado: Hello, friends, welcome to Experian’s weekly #DataTalk, the show featuring innovative data science leaders from around the world. We’re super excited to have Dr. Djamila Amimer.

She is the CEO and founder of Mind Senses Global, which is a management consultancy firm that specializes in artificial intelligence and how artificial intelligence is improving our world. Unfortunately, a lot of the headlines we see talk about how artificial intelligence is taking away jobs — there’s a lot of news like that — and today we’re going to be focusing more on how artificial intelligence is actually improving our world.

Before we get started, Djamila, can you share with us your journey, what led you to begin working with AI?

Djamila Amimer: My journey goes back to my Ph.D. studies. I’ve done a Ph.D. in a combination of operations research and economics. Operations research is basically applied mathematics. So what I have studied is a combination of how to use physiologic genetic algorithms neural network to improve decision-making in the oil and gas sector.
The maths behind artificial intelligence — that’s really my background, that’s what I have studied in the past. That was my starting point, in terms of data science.

Mike Delgado: What aspects of data science really fascinated you from the beginning?

Djamila Amimer: Obviously as a data scientist you have different skills and you deal with different aspects, but for me it was the problem-solving aspect.

I love solving problems. I love having a go at challenging problems and trying to fix them. So it was the problem-solving that got me into data science.

Mike Delgado: That’s great. Do you have any stories you can share, maybe back in college — some challenges, some problems you focused on that you really enjoyed working out?

Djamila Amimer: I’ll give you some examples from my working experience in the oil and gas sector.
Problem-solving skills really can be applied in a variety of ways. I have a number of commercial and business development rolls. One of them was in shipping, looking at liquefied natural gas projects, how to take the gas, liquefy it, ship it to the end port and then gasify it again.

Part of the problem was: How do you come up with an optimal fleet size? What should be the number of ships used to ship the gas, and what is the size of the ships? That’s one of the issues.

Another problem that I have been asked to work on — Whenever you do business development, you consider, for example, market exit or market entry. So you go into a new country, or you are already operating in a country, but you are not sure in terms of performance, profitability. So you have to take a lot of aspects into account. You look at market segmentation, you look at the opportunities to grow the business, but you also look at the risks. What is the risk to the business from a legal perspective, a financial perspective, even climate change? So how you take all those different aspects and make sense from a business context.

And I have the privilege to use problem-solving skills in different facets.

Mike Delgado: That’s awesome. What sorts of data sets are you working with? How large are these data sets?

Djamila Amimer: As you mentioned, I am the current CEO of Mind Senses Global, which is a management consultancy that specializes mainly in artificial intelligence. The mission of the company is really to help businesses in their AI journey. That journey has the two extremes — you get one extreme, the beginner client who has absolutely no clue on AI, so we offer AI education, trying to help build the tools for them. And then the other extreme, the sophisticated clients. That’s the type of client who already has their own internal data scientist and teams that may have their own algorithms.
For that type of client, we offer how to develop an AI strategy, how to gain competitive advantage and how to improve that current algorithm.

But going back to the question in terms of the data and the size of the data, it all depends what type of clients you are dealing with. Is it a small client? Is it a big client? Obviously, there is no straightforward correlation. The bigger you are, the more data you have, or the longer you have been in the business, the more data, because you have longer to accumulate and aggregate data. But I guess it all depends on the problem you’re trying to solve. For some of the problems, you can have huge data set, and it’s fine. For some of the other problems, the data is lacking. You just have to work with what you have.
Sorry, it’s not a straightforward answer. It all depends on the problem you are trying to solve, but here we are trying to deal with different sizes of data.

Mike Delgado: That’s all so cool that this is what you’re focused on, helping companies develop their AI strategies. Can you talk a little bit about the companies out there that don’t know what to do with AI? They approach you and they say, “Djamila, we have all this data. How can AI help us improve our business?” How do you start those conversations with them? Especially for those who don’t have an educational background in data science.

Djamila Amimer: As you know, Mike, there is a lot of hype around AI. AI now is in the news, on TV, in the newspaper, so a lot of people have heard of AI, but not everyone understands what AI is. You get a mix of clients. You get the clients who have heard about AI but don’t necessarily understand what it is, what it can do for the business. Then you have all the types of clients who have a better knowledge in terms of the discipline itself.

I guess so far, kind of how the business works. Some of it has been through networks, some of it going to conferences, trying to raise awareness. I have been giving some talks to groups and organizations, especially with startups, those really keen to understand how AI or machine learning can help them in their journey.
But so far it’s a mixture of word of mouth, advertising, some curious clients wanted to know a little bit more, and so on.

Mike Delgado: That’s awesome. Today we are talking about how AI is improving our world. I would love for you to share some of your favorite examples of how artificial intelligence is being used for good.

Djamila Amimer: This is really a topic that I am passionate about, because obviously there’s a lot of hype around AI, some positive, some not so positive, some on the negative side. So it’s always good to talk about the acronyms for it. The way people refer to it is “AI for good,” which is a good starting point for the discussion.

Mike Delgado: I like that.

Djamila Amimer: Obviously AI has lot of potential, especially when it is used for good, and it has definitely been applied in different applications, different businesses and across the globe. So I’m only going to go through a few examples that have more substance in terms of what we mean by “AI for good.”

There have been a number of feeds where AI has been used for good. One of them is education. I know, Mike, you are based in the United States, so there is a new U.S. platform, Newton, and basically it’s an adaptive learning platform. So the learning is being adapted using machine learning to match with the student [crosstalk 00:09:59] and the student performance.
So the more the student is skilled, the performance is higher, the content becomes more difficult or it increases in the level of expertise. If the performance of the student is lower, or has some difficulties in the learning, then the content gets adjusted to match the expectation and the speed of learning. That’s one example of AI used for good in education.

Mike Delgado: That’s a great example. It definitely would help out in schools so that every student gets tailored curriculum, there’s an AI that can adjust for each child, because that’s always difficulty for teachers in the classroom, right?

Djamila Amimer: Oh, yes.

Mike Delgado: You have students at all different levels, they learn differently, and AI being used in this way can adapt to each student and really provide custom, personalized training and help.

Djamila Amimer: Definitely. And if used correctly and in collaboration with the teaching system, with the teachers and the teachers’ assistants, that will also help them balance their efforts.

They can put more efforts into the students who are behind, and not less of effort, but more targeted content, for the students who are persevering and doing well. So it’s really beneficial to have in education.

Mike Delgado: What is the name of this company that’s doing this?

Djamila Amimer: The platform is called Newton. It’s a U.S. platform.

Mike Delgado: I’m definitely going to check that out. That’s very cool. Is it currently being rolled out and tested in different schools?

Djamila Amimer: I’m not sure how widespread it is, but it’s definitely being tested and operational in some of the U.S. schools.

Mike Delgado: That’s wonderful. I think it’ll be very helpful to so many. I’m interested to see the results of this test, but I love the idea of helping teachers and helping students. That’s a great use case.

Djamila Amimer: Yeah. So that’s one side of the business. There are other examples in other areas. One of them is agriculture, and maybe I’m little bit biased — this is another U.S. example, a U.S. company. This is in agriculture, and the company is called The Plenty.

It’s a new startup that has attracted a lot of investments. Over the last couple of years, there has been more than $200 million spent in this startup.

Mike Delgado: Oh, wow.

Djamila Amimer: So this startup uses IUT, so eternital things, so they use sensors and they use machine learning to have vertical agriculture. So by sensing the climatical condition around the plants, they can guess when the plants need nutrients and need water and need whatever a plant needs, but they can tailor that perfectly around the plant, and then they can increase the productivity and the yield of the agriculture.

It’s now being piloted and run in the United States, but if it goes global, especially in some of the deprived areas, like in Africa where water is scarce, where agriculture is struggling, it could open huge door for a lot of people.

Mike Delgado: That’s wonderful. Being able to provide these farmers with data — maybe even types of seed to be planting.

Djamila Amimer: Yeah.

Mike Delgado: What would provide them the most revenue based on the farming communities there and what they can sell. That’s fascinating. I’m just thinking about all the sensors, and I’m also curious about the cost to do that. But like you said, that is a wonderful thing to be able to do, to be able to help the farming community, especially in emerging countries that need a lot of help.

Djamila Amimer: Obviously, it’s a business proposition, so I’m sure there is monetary value; otherwise this kind of setup wouldn’t even see the light of day. The business cash is there. I don’t know all the details in terms of monetary value, but I believe it adds value from a business context too. So it’s not only good for people; it’s also good for the business.

Mike Delgado: Two purposes, that’s great. So you’ve shared how AI is helping to improve our educational system with Newton. AI being used in agriculture to help farmers produce great crops — using the data to help make informed decisions on when to water, on the soil and seeds. Do you have any other favorite AI use cases?

Djamila Amimer: There are plenty of other examples. I’ll touch briefly on it, and then depending the feed we can tread a little bit more on one of them. This is continuing with the agriculture aspect, but it’s more than agriculture. We’re talking about forest monitoring. How to protect forests.

Forests are a scarce commodity, and to help protect the environment, it’s crucial that we have a better knowledge around that. This is an example outside the United States, just to show that it’s global AI for good; it’s not just for some countries.
Machine learning has been used to predict forest loss. They look at the landscape, they look at the rain forecasts and some climate decondensation, the way the landscape changes in time. And they were able to predict where there could be some forest laws for Congo and the Amazon.

So this is a very useful tool. Using AI and machine learning, you can predict where we may see some issues in the future. That give us a purpose as a society, as people, as businesses, as companies, to do something about it now. With the hope that we could divert from that future.

That’s another example that’s been used. Another one — going back to our health as people, our mental health. We are going back to the United States again. There’s another U.S. company where they have used AI and machine learning for suicide prevention.

I’m just looking at my notes for the right name for the organization. It’s a nonprofit organization called the Crisis Text Line, which is based in New York City.

What they have used — obviously they have tried to monitor the texts and the languages used by the clients, the people who call this hotline. They use machine learning to try to detect which crucial words that, if used by the person on the phone to lead the conversation, may give signals that they may commit suicide in the future. That has been helpful in terms of trying to prevent those suicides that were happening before, because obviously this is a dreadful illness.

When we talk about mental health — This is another way machine learning and AI have been used. Jumping to a completely different area: When you talk about disasters, for example, AI and particularly machine learning have been used to try to predict earthquake disasters, and that hopefully will give the population a longer time to react to the event and take preventive measures.

Mike Delgado: I was just talking with somebody a couple of weeks ago, and he was sharing with me that they have sensors around volcanoes to help detect when a volcano may erupt and provide the data to the communities nearby to give them more time to escape or get to a safer area. They have sensors all around these various volcanoes, and like you were saying, all that is helping to predict if some sort of threat is going to happen.

Djamila Amimer: Definitely. When you think about the potential positive contribution to society, you think, “Ah, it makes sense in applying AI for good.” There are plenty of other examples. I’m not sure whether we will be talking about this later, but obviously healthcare is one of the older sectors where AI has made a significant progress in that. But just for the audience who may not know about this, they’re trying to use AI to help with the United Nations’ sustainable goals. There have been 17 goals developed by the U.N., and there is an Xprize that has been sponsored by IBM Watson, with the view of using AI to help achieve those goals.

I really encourage the audience to go and look, enter the website and look at this Xprize. They will find a lot more information about how AI has been used for good.

Mike Delgado: I took a look at the website and definitely encourage the listeners to check out what the United Nations is doing around using data for good, and all the different initiatives that they have going around the world and their partnership with IBM Watson, which is brilliant. They’re doing some really fascinating work.

We had a guest on a couple months ago, Neil Sahota, who is actually on the committee, and he talked a little bit about …

Djamila Amimer: How fantastic.

Mike Delgado: … the work that they’re doing. It’s brilliant, and it’s so good to see this cause as we’re talking about how oftentimes when we look in the news about artificial intelligence, it’s the terminator stories, the rogue robots, the mass job losses. But then we don’t see what you’re talking about, the positive impact of AI, the new job creation. And how AI is going to be helping our world, just through, like you said, our education system, farming, agriculture, healthcare.
I saw a story recently about how with machine learning, they’re able to help detect cancer faster than radiologists could, just because they could scan much faster and find things quicker.

Djamila Amimer: Talking about cancer … we need some statistics, so it’s not just talk. There are really good studies in this area. In terms of cancer detection, and in terms of the use of AI and in particular machine learning — when you look at skin cancer like melanomas, there has been a lot of progress in that way. If you compare the accuracy rate between the machine and the human doctor, it’s 95 percent for the machine to 86 percent for the doctor.

Mike Delgado: Oh, wow.

Djamila Amimer: In terms of the colon rectal cancer, it’s detecting the early stage. It’s a very difficult cancer to treat if it’s not detected at the early stage. With machine learning, the accuracy rate is around 86 percent.
Another type of cancer, breast cancer …. they lack a lot of research. So, for example, one of the studies that is being currently conducted is DeepMind and Imperial College looking at how AI can improve the accuracy rate.

I’m not saying it’s all jolly good and AI is the answer to everything, but it’s definitely progressing and helping, making those significant steps in terms of improving cancer detection. Hopefully that will give a better likelihood in terms of treatment and survival rate and so on.

Mike Delgado: All the different companies you’ve mentioned, I’m going to put them up on our Experian blog. For people who want to check it out, the short URL is ex.pn/datatalk53. You’ll get links to learn more about Djamila, her work in AI, and also the different companies and startups she mentioned that are doing amazing work to help improve our world.
I’m definitely going to check them out because they sound so cool, the ones you mentioned. Djamila, there are a lot of people in our [inaudible 00:25:36] community who are looking for advice on how to get started. Maybe they went to school, they did their undergraduate or graduate degrees in statistics or mathematics, sciences, and they want to go into data science. But getting that first job can be very, very difficult. Can you share any advice to the young data scientists who are out there looking to get work?

Djamila Amimer: I think that’s a good point, Mike. It’s always difficult to make that first step. Obviously, there is the preparation you mentioned. In terms of the skills, having the right knowledge in terms of the statistics, mathematics, maybe also coding … different aspects of what is a good tool or toolkit for a data scientist.

Making that first step into the business can sometimes be challenging, so I encourage everyone to persevere. Don’t be afraid of rejection, because probably you will be rejected a lot of times before you get your first role. Please don’t let rejection stop you. And then just persevering, networking, trying to advertise your skills. Now with social media and digital skills, it’s so easy for people to blog about data science. Try to do your own little projects. You don’t have to work for a company.

Data is available. There are a lot of websites … if you are passionate about a certain topic, go research it, try to dominate data, do a project about it, and then you can put it in guitar, kind of advertise it. That could be one way to get your foot in to the market. Also going to the job fairs, because data scientists are needed in this new, emerging AI market that is growing. So don’t be afraid of going out to the fairs, socializing. Reach out to some of your connections, people on LinkedIn. You will be amazed how people want to help you.

In terms of the skills, having the right foundation in terms of the maths and the coding is important, but what I think is the most important one, which has really helped me in my career so far, is the problem-solving aspect of it. Challenging the status quo and being able to ask why, what and all those questions. Don’t just run and do the first thing that someone has asked you to. Try to understand what is behind it, how this is going to improve the business, what is the contribution, what is the bottom line. I think it’s doing this kind of thing that will improve the business acumen of the data scientist, and then they will be more in demand, they will be more useful, because they contributed to the bottom line.

Mike Delgado: I love that answer. In conversation, how would you talk to a leader in an organization that is looking to bring on a data science team, looking to hire a data scientist for the first time? What should these leaders be looking for when interviewing a data scientist?

Djamila Amimer: In terms of first skills, again, I think as a minimum that person will have to have the basis. I don’t like to repeat myself, but going back to the maths and science and maybe the coding. But what would be really useful is whether that person is a good fit within the culture of that organization. It’s going back to stakeholder management. Is that person able to communicate in a clear, concise way? Because you could be a very talented data scientist, a genius in terms of the technical bit of it, but if you cannot then communicate your vision to others, this is a big issue. So I feel like it’s a mixture of technical background and the softer skills. Communications, stakeholder management, are they able to lead the team — it’s kind of like a balance of the two.

Mike Delgado: I think you’re right. The communication is going to be key. If you can’t communicate, you’re not going to get buy-in. It’s going to be very, very hard to move products forward if you can’t communicate your goals. Like you said, you can have all the math and the sciences behind you, but if you can’t communicate, lead, there’s going to be some trouble.
Now, along those lines, a lot of people have the hard skills, they have the technical background, but these softer skills, communication … Talk to those data scientists who struggle with that.

They’re like, “I’m listening to you, and I did the stats, I did the maths, I did the coding, I know all the hard skills, but with the softer skills, I’m having trouble.” What advice would you give those people?

Djamila Amimer: I guess it’s just trying to be open, because I know that they have some difficulty in acquiring some skills, and if as a person you feel that you have a gap, I’m sure cannot fear whatever gap you have. That is an answer to it. That is a training course online, a face-to-face type of training.

For example, if you’re looking in terms of training, they like a lot of training in terms of what they call influencing skills, team-building activities. There are courses and training if you are structured and that’s the way that you learn. That’s fine, but I don’t really buy in to a person not having any of those skills. Because as individuals we all live in a society, we have our own families and friends and social activity.

We interact with each other, so we all have, to some extent, some skills.
I don’t believe the argument that someone doesn’t have that. Maybe they need to find it further, yeah, that’s fine. But just be open-minded. Don’t only look into the technical bit and say, “I don’t have it.” I’m sure you have it. You just need to work on it to bring it to the surface.

Mike Delgado: Love that. So the last question for the young data scientists that are out there: What programing languages would you suggest that they focus on? I know there’s a lot of hype around Python, right now.

Djamila Amimer: Yeah.

Mike Delgado: I just want to get your thoughts. If you’re going to learn a coding language, which one would you say would be a good start?

Djamila Amimer: That’s a very difficult question to answer. I used ClassPass because Dispiton was not available at that time, but obviously coding and the languages have evolved since then. I think that the language doesn’t really matter. What matters is really understanding the problem. Having some logic structure in terms of how you go about the problem and some mechanisms in terms of testing the results and feedback, repeat themselves, how you adjust it. In terms of the language, obviously Python has gained some momentum and some publicity over the last couple of years.

If you’re a new data scientist, just starting in the field, why not start with Python? It’s kind of like a video. Another one is R, obviously a lot of the had them mixed, on the research, still using R. But just speak up whatever [inaudible 00:34:48] you think is the most simple and clear and the best fit for you. Don’t get stuck on the language itself. Because what you find out is, once you learn the language, you can very easily converse in the other languages, because even though some of the lines or syntax might change, most of them have some common logic. So don’t be afraid of starting with your own language and then moving to another one.

Mike Delgado: That’s great.

Djamila Amimer: Sorry, no favorite.

Mike Delgado: No, no. I love that. I think I love that answer. It’s more about thinking about the problem and then finding a language that works for you and just learning it. Don’t get so caught up on “I have to learn this, I have to learn that.” Just learn structure, learn the basis of coding; you learn one language and it translates into others. I think that’s really good advice.
Before we go, where can everyone learn more about you and the work that you’re doing?

Djamila Amimer: As you said, Mike, I’m currently heading Mind Senses Global, which is a management consultancy that specializes mainly in AI. I have a website for the company, www.mindsenses.co.uk.

Mike Delgado: Great.

Djamila Amimer: If you would like to know more about the company, what we do and, you know, me, go to the website, which describes how to contact us. Please feel free to contact me. You can also follow me on LinkedIn. I’m happy to help the data science community. And Mike, I have a lot of appreciation for the work we’re doing through your podcast, because that’s one of the tools to help with AI education and just to show that we are all a community, we’re all here to help each other, to bounce ideas off each other. So I’m very happy to be contacted if people wish to do so.

Mike Delgado: Wonderful. So if you’d like to learn more about mindsenses.co.uk or Djamila and her LinkedIn profile and the different companies that she referenced, a lot of the startups, you can find them all on the Experian blog. The short URL is ex.pn/datatalk53. Djamila, thank you so much for being our guest today. It was an honor having you, and hopefully we can have you back soon.

Djamila Amimer: Thank you. I’m very pleased to have participated, so thanks for the invitation.

Mike Delgado: Thank you. Take care!

About Dr. Djamila Amimer

Dr. Djamila Amimer is the CEO and Founder of Mind Senses Global – a management consultancy specializing in artificial intelligence. Djamila received her PhD in Energy Economics and Operations Research from the University of Dundee.

Check out our upcoming data science live video chats.