row hits. Thank you very much for this opportunity as a great pleasure to be here. I grew up in the small city things that in China, and after I graduated from high school, I went to Beijing for my under graduate study, and that was the time when he calmness really started to take off in China. And I was very fascinated by this new way off, not bang and more generally, how technologies can chose one businesses and our daily lives. And so, after graduating from college, I came to the United States and pursue my kids. She started in information systems and management at Carnegie Mellon University, and as I learned more about this field, I realized that there were a lot of open questions related to you. That strategic use of to knowledge is in the business contrast. And ah, my PST species is focused on the use off social technologies such as crowdsourcing platforms, so should lead it performs and so on to engage stakeholders and customers and business processes. And after I graduated from Carnegie Mellon, I first worked at the University of Michigan 14 years as assistant professor aan den. I moved back to Carnegie Mellon 18 2018 and I'm now a factor member at Tipper School of Business, Carnegie Mellon University.

have two main research interests. All the 1st 1 is platforms, and the 2nd 1 is ours. Um uh, So my research platforms, uh, focuses on that in the effective design. Sorry. I hear some I'm sorry. Could I restart this weird? So I stopped recording onI can still hear you.right. I can't hear you now or you may not be speaking.jerk.I have two main research interests the first oneness platform and the 2nd 1 is algorithms and my research on platforms. Fergus is on the effective design and policies toward platforms. Eso, for example. In some of my recent research on, uh, crowd sourcing platforms, my co authors and I have looked at the roll off information such as the in progress, performance, feedback and problem specification in three all come off Carl sourcing contests. Eso, for example. We show that by providing performance three back, for example, you can give ah some start ratings off the excel to the existing submissions. By providing such before mystery back, the firm can actually save a significant amount off the award that, um, it has to pay. Ah, well, intending the quality off the submissions and we also show that it's actually I'm not the best for the firm to provide performance Three bats through all. The context is actually better for the firm to provide performance free by only in the later stage off their crowdsourcing contest s. So this is a concrete example off my research in, uh, in the area off platforms. Ah, and my research, um, in the area off algorithms focuses on the social and economic implications off artificial intelligence and machine learning algorithms. And, um, I look at those issues through an economic lens and highlight importance of considering agents incentives and their strategic behaviour. And so, for example, in one of my working papers, my Coulter's and I look at the economic implications off hours made transparency. Ah, so these days there have been increasing calls for algorithmic transparency off to improve the accountability off algorithms. However, many firms are still very reluctant to make their algorithms transparent, and one of the main reasons they give is that they are. They worry that if they published their algorithms, users will will game the system. Uh, so in this particular study, we consider ah, hiring contest, where a firm uses an algorithm to distinguish high type applicants from the low top applicants, and the Alba Zone takes account. Ah, set all features, and some of them are coastal features, and some of them are correlation of features and an example off. A coastal feature is education. So education. If you improve education, then, uh, the applicants productivity will also improve. Oh, an example of a correlation all feature is whether the applicant is wearing crosses or knelt. Ah, so, uh, this future off wearing glasses that can be correlated with three applicants Type. However, just by simply pulling on your glasses, you cannot really improve your productivity. So we actually show that in this contest when the company published published their l grizzle Um, indeed, people will game the system by, you know, submitting a photo off them wearing glasses. So as the result of this feature will drop out of the Al Gore's own because you don't really learn anything from this particular feature. However, we also show that under certain conditions, high cut individuals they will have stronger incentive. Teoh improve on their coastal feature, basically improve using the same example. Basically, they have stronger incentive to improve their education. And we also show that under certain conditions, the altruism will actually be able to better separate the height of individuals from the low type individual, and also on top of that, because people are improving on their call, so feature, their productivity will also improve as a result. And this will also benefit the firm so they take away from this study is that it is not necessarily true that when a company publishes there, Alvarez, um, when the users are gaming the system, the company will be worse off. So this is the more complete example off my research in the area off algorithms.

when selecting research ideas. The first thing I will consider is whether the idea has any practical relevance and whether it can provide managerial implications for practitioners. Ah, this is very important because we are business school researchers, and the goal of our research is to help improve business practices on. The second thing is that this should be something that I'm interested in and curious about, because the more passion I have about the projects of the better quality I can deliver. And the third consideration is, I guess, um, a feasibility. This is particularly important for empirical projects. Um, before we commit Teoh a project, we really need to assess whether the available data is sufficient to answer the questions we having hand.