 My name is Alexis Korb, and I will be presenting our paper on Beyond the Cizarre Corner Bound Best Possible Wiretap Coding via Obdication. This is joint work with Ivala Shai, Paul Liu, and Amit Saha. We first defined the Wiretap channel, which was first introduced by Winer in 1975. In this model, Alice wishes to send a message M to Bob by sending an encoding of M across Bob's channel, channel B. However, there exists an eavesdropper eave who receives the encoding of M across a separate eavesdropping channel, channel E. Alice's goal is to encode the message in such a way that Bob can decode it, but Eave cannot. In this work, we considered the case for channel B and channel E are both discrete memory list channels. By this, we mean that each channel can be viewed as a function which maps each input symbol to each output symbol with some fixed probability to turn by the channel. Inputs sent into the channel are independently mapped to output symbols based on these probabilities. We are also in the non-interactive setting, meaning that Bob cannot send messages back to Alice, and we assume that Alice and Bob have no prepared shared secrets ahead of time. To build intuition for this problem, we begin by showing a simple impossibility result. If each channel can be used to perfectly simulate Bob's channel, the secure Wiretap coding schemes must be impossible, since Eave can essentially learn everything that Bob can learn. Thus, Bob would have no advantage over Eave, and we cannot send messages to Bob without Eave learning them. We can characterize this by saying that channel B is a degradation of channel E with their existing other channel, channel S, such that channel B can be written as the concatenation of channel E with channel S. Thus, whenever this is the case, we cannot build a secure Wiretap coding scheme. In light of this, we ask whether it is possible to create a secure Wiretap coding scheme whenever this is not the case. More precisely, can we create a Wiretap coding scheme whenever channel B is not a degradation of channel E? Unfortunately, in the information-theoretic setting, this answer is no. Cesar and Korner showed that Wiretap coding schemes are possible if and only if channel E is not less noisy than channel B. This is a stronger requirement than degradation. To illustrate this, consider the scenario where Bob's channel is a binary symmetric channel, which flips each bit with some probability P, and where Eave's channel is a binary erasure channel, which erases each bit with some probability Epsilon. In this case, the entire blue area of the graph on the right represents the set of channels that satisfy the less noisy requirement, and thus cannot be used to build a secure Wiretap coding scheme in the information-theoretic setting, but that are not degraded with respect to each other. As an example of a channel in this blue region, consider the case where Bob's channel is a BSE with flip probability 10%. Observe that if Eave's erasure channel has an erasure probability of 20%, then Eave can perfectly simulate Bob's channel by simply flipping each erased bit to a random value. This would yield a string with approximately 10% bit flips just like the output of Bob's channel. Thus, a 10% BSE is a degradation of a 20% BSE. However, if Eave's erasure probability increases to 30%, then Eave cannot simulate Bob's channel as the best she can do is to flip each erasure to a random bit, which instead yields a string for approximately 15% bit flips. Thus, the example of a BSE with 10% flip probability is not a degradation of a BEC with 30% erasure probability. Seeing as these channels have been very well characterized in the information-theoretic setting, we can ask a very natural question, which is, can we do better in the computational setting? Historically, computational assumptions have allowed us to get much better results in many areas of crypto, including secure encryption and MPC, among many others. But despite the fact that wiretap channels have been studied for decades since 1975, through our knowledge, there has been no other work that studies feasibility results for these channels in the computational setting. We now return to our original question for this time in the computational setting. We ask, can we create a wiretap coding scheme whenever channel B is not a degradation of channel E? And recall that our simple impossibility holds even in the computational setting. Well, in our work, we show that yes! Assuming a form of secure evasive function obfuscation, we can build wiretap coding schemes if and only if channel B is not a degradation of channel E. This greatly extends the feasibility region compared to the information-theoretic setting, and is in fact the best possible feasibility region that we could hope for. Alternatively, our work can be viewed in the ideal obfuscation model, in which case our scheme is unconditionally secure against unbounded adversaries who can make only polynomially many queries to the obfuscated function. To see how we achieve this, please tune into our crypto talk for more details. Thank you!