 Hello, everyone. I'm Yifan Zhang. Today, I'm happy to introduce our work, tight bonds on the randomness complexity of secure multiparty computation. This is a joint work with V4Goya and Yuval Yishai. In this work, we consider the randomness complexity from a practical perspective, generating high-quality randomness from physical sources is difficult. Therefore, in general, we would like to reduce the randomness required in either an algorithm or a protocol as much as possible. From the theoretical perspective, the study of randomness complexity has led to important developments in computer science such as to the randomness, randomness extraction, and so on. We study the randomness complexity of MPC. We start with a simple setting of MPC that is perfect and semi-insecurity against T-cropy parties and computing the XOR function. Later, I will mention extensions to other functions and MPC models. The randomness complexity of an MPC protocol is measured by the number of random coins towed by all parties during the protocol execution. We allow parties to tow different number of random coins in different executions. The randomness complexity refers to the maximum number of random coins towed in execution for all possible inputs. The problem of randomness complexity of MPC has been studied by a fairly large body of works which cover the directions of both lower bound and upper bound. Almost all of these works consider the same simple model as we do. Our research question is to understand how the randomness complexity for computing XOR grows with the number of crowded parties. The best known result for the upper bound is from KM-97 where Kushilevitz and Mansour constructed a protocol with random complexity out of T-square times log n over T random bits. On the other hand, they also gave a lower bound of omega T random bits. In BDPV-99, the authors obtained a lower bound of omega T-square over MST random bits. Then in JRO-05, the authors showed a lower bound of omega log n random bits for T that is at least two. On one hand, when T is a constant, the upper bound from KM-97 matches the lower bound from JRO-05. On the other hand, when T is very close to n, say T is equal to n minus a constant, then the upper bound from KM-97 matches the lower bound from BDPV-99. However, for general T, even if T is equal to n-half, there is still a quadratic gap between the known upper bound and the lower bound. Our first result shows that computing XOR requires at least omega T-square random bits, which matches the upper bound from KM-97 up to a logarithmic factor. We show that the same lower bound applies for arbitrary symmetric boolean functions, such as the n function and the majority function. Our second result constructs an explicit protocol for XOR with order of T-square times log-square and random bits. This is different from the construction in KM-97, which relies on an explicit combinatorial object. Our upper bound matches our lower bound up to a polylogic factor, and we extend our results to arbitrary symmetric boolean functions. Regarding our techniques for lower bound, we connect the randomness complexity to the number of parties' views, and then to the size of the codeword space of T-preventing codings. Our result is obtained by analyzing the size of the codeword space of T-preventing coding schemes. For our upper bound, we give an explicit construction for the combinatorial object in KM-97. Besides our main results, we also show that running T-xOR functions require tilde of order of T-square random bits, which means that each execution only requires tilde of order of T random bits. When allowing helper parties, we give an explicit protocol for general circuits with order of T-square times log-C random bits, where C is the circuit size. For more details about our techniques, please refer to the full video and our paper on E-print. Thank you.