 Hi, I'm Paul, and this is Efficient Nizix from LWE via polynomial reconstruction and MPC in the head, and this is joint work with Riddhi Ghoshal and Amit Sahai. Our starting point is Nizix for all of NP from LWE, which was largely due to CCH-plus-19 and PS-19. Prior to our work, all known Nizix arguments for NP from LWE considered instantiating the Fiat-Shamir paradigm on a parallel repetition of a public coin on a verified zero-knowledge interactive proof. If we have a language L and NP, we can take a statement X and L, compile it via a carb reduction to some instance in a NP complete language such as graph Hamiltonicity, for which CCH-plus-19 and PS-19 give us a Nizix argument in the CRS model based on the hardness of LWE. And this uses some underlying protocol for Hamiltonicity such as FLS-90. The work of HLR-21 extends this to any commit an open protocol such as the three-coloring protocol given by GMW86. In all these protocols, there is a large proof size due to parallel repetition which is done to reduce the soundness error down to negligible. Moreover, this cart production is in general computationally expensive. In our work, we give an efficient base Nizix construction for NP from LWE without parallel repetition in cart productions. We do this by using the NPC in the head paradigm which was first given by ICOS-07. In particular, this allows us to translate work done on efficient, perfectly robust NPC protocols to our setting of efficient Nizix from LWE. Our main informal theorem statement is that assuming the hardness of LWE, there exists Nizix with computational soundness for all of NP, whose proof size is O of C plus Q times depth of C, plus poly K field elements and some finite field F, where K is the security parameter, Q is quasi-linear in K, the size of the field is larger than 2 times Q, and C is an arithmetic circuit for the NP verification function for some NP language L. Here what we mean by a base Nizix construction is that the work of GGI plus 15 shows that we can use fully homomorphic encryption to bootstrap an underlying Nizix protocol to a Nizix protocol whose proof size is just linear in the size of the witness plus poly K bits. In particular, we can view our main theorem as providing a more efficient underlying Nizix from the hardness of LWE. As an overview, our technique builds off of HLR-21's coding theoretic approach to instantiating Fiat-Chemir. They show how to use a list-recoverable error correcting code to address the case of exponentially many bad challenges. There, the block size of this list-recoverable error correcting code determines the efficiency of the proof. They use a Parvavarash Vardy code concatenated with a single random code and achieve a block size of O of K to the one plus epsilon for any positive small constant epsilon. One may ask, can we generically apply this technique to the setting of MPC in the head? In fact, we can. However, general list recovery does not take advantage of the special structure present in the MPC in the head setting which we observe. In fact, we will show that this yields less efficient proofs. And in our work, we note that the bad challenge set structure present in a modification of the ICO-07 protocol only needs a strictly weaker notion of list recovery which we term recurrent list recovery. Therefore, we can use qualitatively simpler codes to instantiate recurrent list recovery, namely Reed-Solomon codes concatenated with multiple random codes, and directly use the polynomial reconstruction algorithm given by Sudan-97 and Guru Swami Sudan-98 to achieve an improved block size of quasi-linear in a security parameter K. For more details, please come see our talk. Thank you.