Unique Contributions

Artificial intelligence and the future of law

RELX Season 2 Episode 5

Gender balance in the world of Artificial Intelligence is work in progress. But female role models do exist. In this episode, YS Chi speaks with one of them. Min Chen is a computer scientist leading a team of around 250 technologists and software engineers focused on applying machine learning, natural language processing and other advanced technologies in the legal industry.

Min was one of the few female students who studied computer science in Shanghai in the 90s, and the only one from her class to pursue a career in the industry. She tells us where she found the inspiration to ‘stick with it’ and overcome adversity. She talks about the profound technological transformation taking place in the legal industry and gives us a flavour of how technology is changing the way lawyers work. Here are the reflections of a female technologist rocking the fascinating world of AI and deep learning.   

Min Chen is chief technology officer of LexisNexis Legal & Professional, part of RELX, for Asia Pacific and Global Search.

This podcast is brought to you by RELX.

YS Chi:

The Unique Contributions podcast is brought to you by RELX. Find out more about us by visiting RELX.com.

Min Chen:

I want to make it very short and simple as you have heard my story earlier. I want to quote my math teachers remark. Find things you're really passionate about and stick to it, and don't give up.

YS Chi:

Hello, and welcome to our second series of unique contributions, a RELX podcast where we bring you closer to some of the most interesting people from around our business. I'm YS Chi and I'll be exploring with my guests some of the big issues that matter to society, how they are making a difference, and what brought them to where they are today. Hello, and welcome back to our listeners. Our guest today is Min Chen, Chief Technology Officer of LexisNexis Legal & Professional for Asia Pacific and Global Search. Based in Shanghai, Min runs large teams of technologists and software engineers to deliver solutions to clients in the legal industry. The legal industry is going through a profound technological transformation. I'll be asking Min what this means for lawyers, and how the new generation called Gen Z is shaping the future of law. Along the way, we'll also hear about what it's like to be a woman working in the field of technology and data science. So hello Min, thank you for joining us even though it's a night time in Shanghai and the weekend.

Min Chen:

Hi YS, I'm honoured to be the guest.

YS Chi:

I would like to kick off by talking about your actual day to day work at LexisNexis. Over a decade ago, you joined a team with less than 10 people to support the China business. Then today as CTO of LexisNexis, Asia Pacific and Global Search, you oversee a team of about 250 technologists and software engineers spread across everywhere from Shanghai to Australia, US and India, etc. The projects you work on are scattered all around the world. This reflects a profound technological transformation that is taking place in the legal industry it sounds like. So please tell us a little more about that.

Min Chen:

Right. I think the legal industry has been relatively conservative to adopting new technologies in the past century because legal systems are always jurisdiction specific, and the evolvement of law is also relatively slow. Having said that, I do notice that rising demand of embracing and seeking analytics and automation by our customers has been pushing changes on legal industry through advanced technology, particularly on cloud based solution and artificial intelligence. The pandemic accelerates this change a bit because it pushes new ways of gathering, sharing and dealing with data. For example, law firms want to automate those repeatable manual tasks to keep their rates competitive, so they can continue to win business. While in-house counsel wants automation to lower the amount of money they have to spend on using outside legal counsels. They will all need a better decision making tool to help them and descriptive or predictive analytics, will play a big part of it. LexisNexis' mission is to help our customers to improve their work efficiency. We do have a big advantage to deliver that mission because we have both data and advanced technology that includes natural language processing, natural language understanding, generation, machine learning, deep learning, computer vision, and so on.

YS Chi:

So Min, at some point in time mindset began to change. Were you there to observe that change in mindset and how did LNLP decide to jump onto that?

Min Chen:

Yeah, I was there. Pretty much we are seeing the recent three to four years. I see those changes and that is why the team has been growing from 10 to 200 at the same period. Legal information is mainly tax based and that's why artificial intelligence plays a critical role to address the issues of natural language understanding and natural language generation.

YS Chi:

This so called'latecomer' to technology transformation is suddenly having a real fast pace. Is it pandemic or is there something else?

Min Chen:

I think it's driven by the customer problem. The pandemic definitely is part of that, as already mentioned. The pandemic actually generates the new ways of thinking, getting the information. But I think it is driven by the customer problem. We have a tough customer problem in the legal domain. Legal information is very often complex, professional and lengthy text based documents. Given that specialty, we need expertise and focus particularly on natural language understanding. I earlier mentioned about that, I'm just very obsessed with that. So natural language understanding as well as machine reading comprehension, having the machine to understand the context of the body of the text. Natural language generation is about competition producing, having the machine to produce the text. Deep learning technology that is evolving faster than ever in the world of AI, will clearly be the major solution to tackle these areas. So when I say it is driven by customer problem, I give you one example here, like auto summary in legal domain. We have this case summary which is critical for legal researchers. So without case summary, a legal researcher may need to read through the whole case before they can be sure whether the case is applicable to their case or not. But generating summary is not free. It highly relies on human efforts, and only well trained legal practitioners can understand the legal domain language. However, there are millions of the case there and getting labour efforts on such work is very challenging, because it's both time and resource consuming. In order to improve the coverage of high quality case summary to our customers, we delivered an automation solution to accelerate the case law document summary process through multiple state of art, deep learning technologies. Given the length of the case law documents and complexity of legal domain, this target task is even more challenging than other summarization tasks in other industries you've seen. That's why deep learning can make a big difference.

YS Chi:

I suppose that the domain expertise of our folks in LexisNexis Legal and Professional really allows us to work with technologists like you and your team to jump that queue quite quickly, doesn't it?

Min Chen:

Yeah, they actually love to work with us because they have the domain knowledge. Every domain knowledge, if you want to improve the coverage, improve the speed, they need to work with a technologist to do all the automations. So basically, they have their insights. Combined with our strong advanced on tech knowledge, these two things combined together can make the magic.

YS Chi:

Now, you've talked very passionately about natural language processing and deep learning. Are there any other technological breakthroughs or changes that you have noticed that has really jumped and accelerated these transformations?

Min Chen:

I would come back to deep learning, because deep learning is a very broad area. It is like artificial intelligence. You have lots of things, but the traditional way to tackle the problem is the traditional machine learning. There's a limitation there. But deep learning actually is a big job to resolve the natural language understanding and natural language generation issue. In recent developments, we launched a new product called Asia Legal Analytics. That product is using the deep learning to do the knowledge extraction. The traditional way we use the natural language processing and traditional machine learning, we can do entity extraction. You can extract lawyers name, law firms name and court names, and just judges name. So that's the traditional entity extraction. Now we have a big job we call the knowledge extraction. Basically, you have this case where we're able to extract a legal issue. What is the legal issue for this case? What is the legal argument for this case? What is the legal principle for this case? That we call knowledge and insight extraction. If you only use the traditional machine learning, you're not able to get the insight extraction. But by having deep learning technologies really helped us to make the difference. We just launched that in February in the Malaysia market. We're going to learn launch that in April for the Hong Kong market. After our initial launch, the customer already feels amazed because in the past they see LexisNexis, we have Lex Machina, we have Ravel contacts. In the US market, they already have this kind of analytics. But now we have this analytics focused on insight extraction. So the customer feels amazed, it really is very successful in the Asia market.

YS Chi:

I'm am equally intrigued. I have a dumb question. Does this work on all languages? Or does it only work on certain global, widely spoken language that are single byte?

Min Chen:

Yeah, I think the overall technology is suitable for all languages. But there's a different approach and techniques we have to take. It's because when we deal with different language, it's basically about token. How do you extract the token? The double byte language is not like English, the sentence is separated by meaningful tokens. So the way you extract the token, in a sentence is different. You just have to get through that part and the rest of the techniques can be leveraged across different languages.

YS Chi:

Is this very expensive to do Min?

Min Chen:

It really depends on the problem you're talking about. So I can't...

YS Chi:

It just sounds so fancy. It sounds so complex, and therefore you have to think it's expensive. But obviously, you do it in a very efficient way.

Min Chen:

Yeah, well once you build a foundation there, because we already have millions and billions of data there. LexisNexis has the data so we already build the foundation. Maybe when you build something from scratch, it takes time. It takes multiple experiments to get there. But once you set up, build up the foundation, you already have that algorithm there, it will be easier to leverage to other regions. To build up your first concept, it might be not very cheap. But if we want to expand a similar idea to other regions, it becomes cheaper and cheaper.

YS Chi:

That is really one of our unique advantage and unique contributions because we have both the content and now the technology. Well, as unique as that is, I find you to be a very unique person Min. Since we have met several years ago, I've been fascinated about your growth and I'd like to turn to that for now for a few minutes. But before we do that, tell us a little bit about how you're handling this, you know, unexpected and very prolonged pandemic.

Min Chen:

Well, I want to show off a little bit YS. We have been completely back to the office for a year now. I'm definitely not, I'm not a big fan of working from home, even before the pandemic. So I do acknowledge we have advanced to technologies to help people meet virtually. For example, a few weeks ago, I was demoed by the team for an AR and VR app to allow engineers to do pair programming and code review remotely. By the way, pair programming is where two engineers work together on the same piece of code. Code review is a group of engineers where they get together to review the same piece of code. So you can imagine these kind of activities require a heavy in person experience. Now, there are different types of tools which can support engineers to do the same work virtually. But I think the dynamic and the inspiration of being able to have face to face conversation cannot be completely replicated. There are just a few exquisite nuances that you might not be able to tell them explicitly, but you just know the difference. It's that difference that can make a difference to be creative, to drive innovation in a more efficient and productive way. So once the situation became better in Shanghai, it was only like three to four weeks, there was locked down. So we were all working from home. But when the restrictions started to ease, I brought the entire team back to the office and everyone was super excited to meet everyone in 3D again. Having said that, it seems only the China gets 100% back to office, and the rest of the world still majorly work from home, including our customers that we're serving in Asia Pacific and globally. So to me, the ways of interaction has changed. I think I've grown my remote research skills to the new heights. During the pandemic, I got more chances to talk with customers remotely. That really helps me to learn a few innovative techniques to capture the precise feedback from customers for product validation or discovery through phone calls. That I consider as the most valuable new skills I developed during the special period.

YS Chi:

Always looking at the positive side Min. So, you know, some years ago, it would not have been at all conceivable that I would be doing this kind of conversation with a women technologist. It's been a little over 15 years since you joined LexisNexis. Then you've spent a little more time in the broader tech world before that. You even studied computer science at Shanghai University. Right?

Min Chen:

Right.

YS Chi:

You were one of the very few girls in that area. There must have been a lot of competitive and demanding environment that you had to cope with at the time. Would you like to kind of share some stories about that?

Min Chen:

Yeah, I can try. I think at different stage of my life, I am lucky to always have the unique driver that strengthens my face to stick to technology industry. Back to high school where each student was at the moment to decide what is the future of their major subject in college. My parents want me to follow either of their pasts. My mother is a professor of Chinese literature and my father is the professor of biology. I didn't know what I wanted to be. It was my math teacher who encouraged me to pick computer science. He said, I think that will be a perfect fit for you. Honestly, at that time, I didn't know what computer science meant to me, until I followed his instructions and then went to Shanghai University to major in computer science. Very soon, I found myself obsessed with programming. I could spend hours and hours in the library to write a code just for a better version to beat myself. So it was not the homework that I have to finish. I really like the feeling that I can just write a few lines of code to resolve issues automatically. That's amazing. I remember my scores were always top three in a class, but there was one thing that bothered me. Regardless of my high scores, my professor never picked me to join an external science programme with top schools in Shanghai. At that time, we had quite a few learning and competition programmes with leading universities, like Fudan University, Jiao Tong University, and usually top students in the class got picked. Those were great opportunity to learn from other top students of different schools. So I felt a bit upset that I was not elected. So I called my high school math teacher, I always kept contact with him. I asked him, I said, you told me computer science would be perfect for me, and look, I don't even get the chance to join these programmes. Do you still think I'm suitable to learn the subject? He responded with two things which I'll never forget. He first asked me, do you love computer science? I replied, yes, I love it. So I talked to him about all the stories that I spent a long time to study and to learn, all out of my own interest. Then he said, don't worry about now. Now you don't get chances to join all these programmes in school, but I have faith, you will get lots of chances when you get a job, when you start a career. As long as you love what you're doing and you don't give up. So that really motivated me. I recall in my class, there were 50 students, only eight were female, and out of eight girl students I was the only one to pursue a career in the technology industry after graduation.

YS Chi:

Oh no.

Min Chen:

I was the only one. So my first job was a programmer in China Daily. Then three years later I joined Lenovo being a tech lead for another three years, and then LexisNexis. So 10th of October of this year, will be my 16 years working anniversary. A lot of people ask me why you stay in this company for that long. Obviously there are multiple reasons. In LexisNexis there are enough great challenges for me to tackle by using advanced technology. You heard the story, deep learning, machine learning. My teams and colleagues, they're all great. I'm growing together with them. But I think there's one critical reason I never ignore. Is that in this company I come across a few of my math teacher type of people. They share very similar value and vision which is doing things that you love and don't give up before you give up. So that keeps me motivated in the long run.

YS Chi:

Well, can I ask what the name of your math teacher was, because I'm inspired by his character.

Min Chen:

His name is Xu Quanxiang. He's just like me, a normal person. I don't know a lot about his personal life, but he is very disciplined, very organised and a very fair gentleman. In class he always encouraged me to talk, a lot. So he asked the question, and some students raise their hand. I'm not the student who raised their hand. He will always call out my name, then I can always answer the question. So once I did ask him, why you always call out my name? He replied, he was smiling. He said, because I know you know the answer. You just need to stand up.

YS Chi:

Oh, wow.

Min Chen:

Yeah. So unfortunately he passed away 20 years ago, but the legacy he leaves to me will forever remain.

YS Chi:

Wow, we should all be lucky enough to have teachers like that through our lives.

Min Chen:

Yeah.

YS Chi:

Well, obviously, you've done very well for yourself, for your teams and of course for our company since joining. We're very appreciative of your work and passions. You were named in 2019 as our distinguished technologist. Acompany wide award that only one person wins every year for exemplary leadership, and fundamentally challenging and changing the way we do business. What is more important than that is that you were the first woman to receive that award. What did that mean when you found out that you were the first woman to win that prestigious award?

Min Chen:

I'll will tell you the story. So, when I initially learned this news from my manager, Jeff Reihl, who is the global CTO of LexisNexis. Of course, initially, it was like a surprise. It was announced by Erik Engstrom and Kumsal. They organised a surprise call and they announced it to me. But it was a very short call. Then, when we get off the call, I talked with Jeff. He also highlighted, he said, you're the first woman who won this award. But my first, my instinct reaction was, is that the reason I got selected. Well, to me it's just like the two sides of the coin. I don't want to be overlooked for opportunity because I am a woman. But on the other hand, I also don't want to be recognised because I am a woman. Of course, Jeff said it has nothing to do with your gender identity. I was recognised because of my contribution and delivery to our company. That I have been building up the culture of customer driven innovation amongst the engineering team and developing high quality products that differentiate us in the Asia Pacific market, and being part of the solutions to improve customers NPS, the net promoter score. So I consider this award as a recognition of me being able to make an impact on our customers and business. I will strive to extend that impact to much broader customers so it's not just Asia Pacific but also globally. I think I'm moving toward that goal closer and closer every day.

YS Chi:

You sure are. I remember sending you a congratulatory note after the award.

Min Chen:

I remember that. Thank you so much for the encouragement YS.

YS Chi:

You are encouraging a lot of girls through your own journey. I'm curious to know, what do you think RELX can do more to improve the pipeline of women coming through to all fields, including STEM where they seem to be not yet in abundance? What advice would you give female technologists who look up to recognise leaders like you?

Min Chen:

Yeah, I think I have already seen lots of forums and panels organised with RELX to motivate and promote women. For example, in general I joined the panel session of RELX Thrive global launch in Asia Pacific. Thrive is a grassroot organisation to bring all female together and provide a portal for them to share thoughts and stories. I know there are many other communities like this within RELX. I think RELX can do more on gender balance, because that could push a real force for driving innovation. To me gender balance doesn't necessarily mean we have to favour a particular gender. It means we have to consciously check whether we have enough difference in the team. When we have different ways of thinking, personality and culture, it can stimulate innovation. That's why it's important. I think we're also changing the way we make hiring and promotion decisions, and to ensure that eligible women are given serious consideration. Last but not least, I do want to highlight, let's not forget how critical it is that our male leaders support all these initiatives to me. The technology industry is still a male dominant area, and that situation I don't foresee will be changed drastically overnight. So our male leaders who are very supportive on consciously improving gender balance in organisation can make a big difference. For example, I know YS you are a great sponsor and advocate on driving diversity and inclusion. In my organisation LexisNexis, Mike Walsh our CEO, and Jeff Reihl the global CTO, and Jamie Buckley global CPO, they're all strong supporters. I often think how great the D&I we have in our organisation. It is majorly depending on how open and how far our male leaders are embracing these ideas and really taking to actions. Finally, in terms of what advice I could give to other female technologists. I want to make it very short and simple, as you have heard my story earlier, because I want to quote my math teachers remark. Find things you're really passionate about, and stick to it and don't give up.

YS Chi:

Yeah, I think that is clearly the theme. I think the dimension of allies, male allies was a key topic when we discussed during the International Women's Day panel. We do seek women to do the work but men ally, I think have the responsibility to provide the air cover so that they can fight on the ground as effectively as they can.

Min Chen:

Yeah, exactly.

YS Chi:

Well, let's jump back to the business world that we left earlier in our conversation. LexisNexis has a number of tech hubs. One is in North Carolina, one in London, and here you are in one in Shanghai. How do you keep such a big team of data scientists, engineers and product managers spread all over different locations? Yet, you're working so well, so motivated, especially through out the pandemic? What are the secret sauce?

Min Chen:

Yeah well, pandemic for me was very short. Early I mentioned I only worked from home for only three to four weeks, and then we were back. I think I talk from a general perspective. There are a lot of initiatives we have been doing to keep the team motivated. For example, we do hacks once or twice a year to bring engineers and product managers together, in order to incubate creative ideas and solutions through non-stop programming for 24 to 48 hours. Eventually each region's winner will join a fun competition, we call the global Shark Tank to present our work to our global CEO, Mike Walsh, and his MCM. Mike and his senior leaders will pick a winner of the winners, and my team are always winning and we have won top prize globally three years in a row. Besides, I want to mention another initiative that I ran for a few years in Asia Pacific, to allow people to work on innovation on regular basis. We call it 'grab the post.' So the idea is we pop problem statements in one page, and post it at the door of my office. These problems are very preliminary ideas from either internal or external customers, which are not yet put into official business case or product development roadmap. Therefore they're treated as a side project. Whoever has interest to take the challenge during their free time can just grab the post and commit to the POC, proof of concept delivery.

YS Chi:

Aha.

Min Chen:

Yeah. That's why we call it grab the post. Since this is not mandatory, and it's a side job, you can imagine in the end, whoever grabbed the post is truly passionate in building great customer experience through leadi ng edge tech knowledge. We do acknowledge the situation, that some of the posts have never been grabbed by anyone. But averagely I did amass, every three to five months we will get one or two challenges rejected by the team with proof of concept delivered and we gave the incentive to the team who grabbed the post. So if you deliver the POC and that idea got invested by our business partner, you will be the person to lead final delivery onto production. You have no idea what being able to bring the idea solution live to customer means to engineers, because it's a great motivation for engineers.

YS Chi:

Oh, yes. That's like being selected in the starting lineup. Well, time is flying so I'm going ask one last question before we close. You have been very good at sticking to what you're passionate about. Obviously because you probably saw things down the road. So what do you see down the road in terms of transfer tech that you expect to see over the medium term? And how are you going to make sure that your team does right on that journey?

Min Chen:

That's a good question. I think our focus and investment priority will always be on those specific technologies that could address customer pain points. So from mapping customer problem to technology perspective. I want to summarise two major trends, two major areas. One is customer's expectation, and quality of service we deliver to them will continue to increase. So that requires more efforts on data science work. The other observation is that a customer would expect technology to create more personalised and tailor made service to meet their specific individual needs. So those technologies that could move sophistication of personalization into the next level, will be hard. Of course, you have to make sure personal data privacy is not violated. So getting understanding and deeper understanding with customer problem as the guideline, as the major driver for us to prioritise our technology, strategy and focus. Basically, we stay close with a lot of communities and channels, both internally and externally. That we could constantly share the most current advanced technologies of different kinds. For example, we always keep a close eye on global external top conference and top papers out of those conference. Particularly in AI industry, of course, in other technology industrys as well. That helps a team to stay current. So it's just not just me, I force people to learn this and that. You have to build a culture there, that culture should be everyone is eager to learn. But that culture is always driven. So in terms of what we want to learn, that should be always driven by what customer problem we want to resolve. So with that, because there's so many technologies, it's very broad, right. There's so many different layers, different levels of the things you have to learn. You cannot learn all of them. So, you we use the customer problem as the guideline for us to decide which area we want to do the deep dive.

YS Chi:

Well, you had a programme called getting outside the building and getting your tech people to listen to customer firsthand?

Min Chen:

Yeah.

YS Chi:

Yeah. That's awesome. Now, I cannot close without asking a follow up question, because you know, tailor-made solution is obviously something that we all want to do. But tailor-made has always been associated with high cost, high, high price. So this technology adoption, is meant to try to make it tailor-made without high costs. Is that right?

Min Chen:

Yeah. Put it this way, I would love to rephrase that into a cost effective way. But actually, it's the same thing. You're right. It's the same thing. We're going to want to do the tailor-made solution in a efficient way, in a cost efficient, effective approach.

YS Chi:

Yeah, I think that we can do what someone has coined mass customization, right?

Min Chen:

Yeah.

YS Chi:

Well, I'm fascinated and I could stay on for three hours talking to you Min. It's always been fun to talk with you. Every time I meet you, you are five steps ahead.

Min Chen:

Thank you

YS Chi:

Thank you so much Min.

Min Chen:

It's also an honour and I also feel pleasure talking with you YS.

YS Chi:

Well, thank you for sharing your stories and insights with us today. I think Min has shared with us not only about the technology of natural language processing or deep learning. But really just about her subtle leadership that is not about just being a pioneer and bringing the entire team along, and making this unique contribution to the legal community. Thank you again. Thank you to our listeners for tuning in. Please don't forget to hit the subscribe button on your podcast app to get the new episodes as soon as they're released. Once again, thanks for listening and please stay well and healthy.