Federated learning has received increasing attention for its ability to collaborative learning without leaking privacy. Promising advances have been achieved under the assumption that participants share the same model structure. However, when participants independently customize their models, models suffer communication barriers, which leads the model heterogeneity problem. Moreover, in real scenarios, the data held by participants is often limited, making the local models trained only on private data present poor performance. Consequently, this paper studies a new challenging problem, namely few-shot model agnostic federated learning, where the local participants design their independent models from their limited private datasets. Considering the scarcity of the private data, we propose to utilize the abundant public available datasets for bridging the gap between local private participants. However, its usage also brings in two problems: inconsistent labels and large domain gap between the public and private datasets. To address these issues, this paper presents a novel framework with two main parts: 1) model agnostic federated learning, it performs public-private communication by unifying the model prediction outputs on the shared public datasets; 2) latent embedding adaptation, it addresses the domain gap with an adversarial learning scheme to discriminate the public and private domains. Together with theoretical generalization bound analysis, comprehensive experiments under various settings have verified our advantage over existing methods. It provides a simple but effective baseline for future advancement. The code is available at https://github.com/WenkeHuang/FSMAFL.

The public cloud data integrity auditing technique is used to check the integrity of cloud data through the Third Party Auditor (TPA). In order to make it more practical, we propose a new paradigm called integrity auditing based on the keyword with sensitive information privacy for encrypted cloud data. This paradigm is designed for one of the most common scenario, that is, the user concerns the integrity of a portion of encrypted cloud files that contain his/her interested keywords. In our proposed scheme, the TPA who is only provided with the encrypted keyword, can audit the integrity of all encrypted cloud files that contain the user’s interested keyword. Meanwhile, the TPA cannot deduce the sensitive information about which files contain the keyword and how many files contain this keyword. These salient features are realized by leveraging a newly proposed Relation Authentication Label (RAL). The RAL can not only authenticate the relation that files contain the queried keyword, but also be used to generate the auditing proof without sensitive information exposure. We give concrete security analysis showing that the proposed scheme satisfies correctness, auditing soundness and sensitive information privacy. We also conduct the detailed experiments to show the efficiency of our scheme.