 Welcome to the top on multi-input attribute-based encryption and predicate encryption. I am Manshu Yadav and this is a joint work with Shweta Agrawal and Chodayamata. Attribute-based encryption is a primitive that, as we know, provides fine grid access, control, or encrypted data where data is encrypted with respect to an attribute X and the keys are generated with respect to quality CAF so that the decryption is possible if and only if FX is equal to one. And in a more general notion of predicate encryption which provides stronger security, the ciphertext hides the attributes as well as long as the adversary does not issue any decrypting key queries in the security. So let us begin with the motivation behind the work. So the starting motivation comes from the fact that thinking of a data to be associated with a single ciphertext may not suffice for many modern world applications where the data related to a single entity can in fact be distributed across different locations. So for example, consider a for the following scenario where there is a medical researcher who wants to study efficacy of certain medicine in patients, let's say over 65 years of age with both COVID and asthma. In such a situation, we would like to use predicate encryption to provide respected access to records of only those patients who are over 65 years of age and have COVID and asthma. However, note that in the real world scenario, records of any patient may be distributed across different departments of the hospital or across different centers. For example, COVID center may be located in a different part of the city. However, we would still want to be able to use predicate encryption to provide controlled access to the data. So basically, what we want is that each department should be able to generate encrypt the ciphertext locally. And then the researcher should be able to use her key on the combination of these independently generated ciphertext to recover the underlying data. And in general, we would like to think of AB or PE in any such scenario where the data is generated at different locations, but is it still correlated so that it forms a single logical group? It's just that the different components are generated in a distributed fashion. So let us look at the related work in determining. In fact, even though this primitive being quite useful, it is largely unexplored. The notion of MIAB was introduced by Rikersky et al. in the context of constructing witness encryption. However, they do not provide any concrete construction. And in the case of MIB, there is no study at all. So let us look at the main difficulty in constructing MIAB. So the main difficulty comes from the need to satisfy two seemingly contradictory requirements at the same time. So on the one hand, we want to be able to generate different components of the ciphertext independently. And on the other hand, we want to be able to combine these ciphertext so as to perform the encryption. And if we look at the existing constructions for ABE, the second requirement needs that the randomness used in these ciphertext components is same. And on the other hand, the first requirement says that these randomness are generated independently. And this contradictory requirement makes it difficult to construct MIAB. So here are the results that we get. Firstly, we formalize the security definition for MIAB and MIB. And then in MIAB regime, we get two ABE for NC1 and heuristic constructions for three ABE for NC1 and two ABE for P. And the main insight in our result comes from the finding that there exists a surprising relation between techniques developed in context of succinct ciphertext policy ABE to a completely unrelated setting of two input to key policy ABE. And we exploit this relation to get the results that we have. And then in the regime of MIB, we give a generic compiler to lift any k-input ABE for a constant k-to-ABE. And then using this compiler on the results that we have for MIAB, we get two ABE for NC1 and heuristic constructions for three ABE for NC1 and two ABE for P. Thank you.