 Hey everyone, this is Jaspal and today I'll be giving an overview of our new work on structure of a private-set intersection. This is a joint work with my co-authors, Gayatri Garipela and Mike Rose Lake at Oregon State University. So, let's begin by looking at the traditional private-set intersection problem. So, there are two parties, Alice and Bob, and they both have a set of discrete points and they are interested in learning the set of points that are common to both their sets. However, there is an additional privacy constraint which states that both parties should learn nothing but just the intersection, which essentially means that Bob does not learn any of the elements of Alice's set outside the common elements O, N, R, and S. So, this traditional problem of PSI has been well studied in the literature, so let's consider a variant of this where Alice and Bob are holding a set of GPS coordinates in and around Santa Barbara and they are interested in learning if they both have GPS locations of any common locations. Now, it's extremely unlikely that Alice and Bob will have the identical GPS coordinates for a location. The more likely case is that they would have GPS coordinates which are pretty nearby. Hence, we can formulate this problem as a problem of first-zpsi where our goal is to learn the points of Alice and Bob which are within some distance delta for some parameter delta. So, one naive way to solve this first-zpsi problem is to redefine Alice's set as all the points contained within a delta radius ball around all its GPS coordinates. And then we essentially run plain PSI between Bob's points and Alice's expanded set. Traditional PSI protocols have communication complexity proportional to each party set. So, in this case Alice's communication would be proportional to the expanded set size which in this case would be union of delta radius balls. However, unlike plain PSI, in this case there is some known structure over Alice's input set. It's a union of delta radius ball. So, question is can we improve the communication complexity of PSI when there is a known structure over one of the party's input sets? So, in this work we introduce a new framework for structure-aware private set intersection which gives a very viable approach for doing semi-honest fuzzy PSI as mentioned before. So, we give very practical symmetric key-based PSI protocols for the case when Alice's input set has some publicly known structure. In our work, we reduce this problem of structure-aware private set intersection to a new variant of function secret sharing which we term as weak FSS. Using this framework, we reduce the PSI communication of Alice to order lambda times FSS share size where lambda is the security parameter and FSS share size is the share size for secret sharing the structured set of Alice. So, this is a considerable improvement over just the carnality of the expanded set in the cases when the FSS share size can be really small for structured sets. Other than the framework, we also give a number of new FSS techniques which allows us to construct really efficient weak FSS for union of geometric balls which has applications for fuzzy PSI. To learn more about our PSI framework and our new FSS techniques, you're welcome to attend my talk at Crypto 2022 or read our paper for a thorough discussion on all of these techniques. Thank you.