Abstract. Recently, opacity has proved to be a promising technique for describing security properties. Much of the work has been couched in terms of Petri nets. Here, we extend the notion of opacity to the model of labelled transition systems and generalise opacity in order to better represent concepts from the work on information flow. In particular, we establish links between opacity and the information flow concepts of anonymity and non-interference such as non-inference. We also investigate ways of verifying opacity when working with Petri nets. Our work is illustrated by an example modelling requirements upon a simple voting system.

Abstract. Recently, opacity has proved to be a promising technique for describing security properties. Much of the work has been couched in terms of Petri nets. Here, we extend the notion of opacity to the model of labelled transition systems and generalise opacity in order to better represent concepts from the work on information flow. In particular, we establish links between opacity and the information flow concepts of anonymity and non-interference such as non-inference. We also investigate ways of verifying opacity when working with Petri nets. Our work is illustrated by an example modelling requirements upon a simple voting system.

We introduce a class of causal video understanding models that aims to improve efficiency of video processing by maximising throughput, minimising latency, and reducing the number of clock cycles. Leveraging operation pipelining and multi-rate clocks, these models perform a minimal amount of computation (e.g. as few as four convolutional layers) for each frame per timestep to produce an output. The models are still very deep, with dozens of such operations being performed but in a pipelined fashion that enables depth-parallel computation. We illustrate the proposed principles by applying them to existing image architectures and analyse their behaviour on two video tasks: action recognition and human keypoint localisation. The results show that a significant degree of parallelism, and implicitly speedup, can be achieved with little loss in performance.

Abstract. In this paper, we extend previous results relating the DolevYao model and the computational model. We add the possibility to exchange keys and consider cryptographic primitives such as signature. This work can be applied to check protocols in the computational model by using automatic verification tools in the formal model.To obtain this result, we introduce a precise definition for security criteria which leads to a nice reduction theorem. The reduction theorem is of interest on its own as it seems to be a powerful tool for proving equivalences between security criteria. Also, the proof of this theorem uses original ideas that seem to be applicable in other situations.

In this paper we identify the (P, Q)-DDH assumption, as an extreme, powerful generalization of the Decisional Diffie-Hellman (DDH) assumption: virtually all previously proposed generalizations of DDH are instances of the (P, Q)-DDH problem. We prove that our generalization is no harder than DDH through a concrete reduction that we show to be rather tight in most practical cases. One important consequence of our result is that it yields significantly simpler security proofs for protocols that use extensions of DDH. We exemplify in the case of several group-key exchange protocols (among others we give an elementary, direct proof for the Burmester-Desmedt protocol). Finally, we use our generalization of DDH to extend the celebrated computational soundness result of Abadi and Rogaway [1] so that it can also handle exponentiation and Diffie-Hellman-like keys. The extension that we propose crucially relies on our generalization and seems hard to achieve through other means.