26838 research outputs found
Sort by
Curriculum design for scalable biologically plausible deep reinforcement learning
Get PDFHumans have a remarkable capacity for learning, yet neuronal learning is constrained to locality in time and space and limited feedback. While neural learning rules have been designed that adhere to these principles and constraints, they exhibit difficulty in scaling to deep networks and complicated datasets. BrainProp is a biologically plausible learning rule, learning from trial-and-error feedback through reinforcement learning, that does generalise to deep networks and achieves good performance on traditional machine learning benchmarks. It does however falter on problems with a large number of output categories, such as the classical ImageNet vision benchmark: while standard BrainProp eventually succeeds, learning is not robust and highly sensitive to hyper-parameter optimisation and proper initialisation. Here, we leverage insights from behavioural science by developing a curriculum that structures how samples are presented to a network to optimise learning. The key features of the curriculum involve progressively introducing new classes to the dataset based on performance metrics, and using a recency bias to protect recently acquired classes. We demonstrate that our curriculum approach makes BrainProp-style learning robust and more rapid, while substantially improving classification accuracy. We also show the curriculum similarly improves performance for networks trained using error-backpropagation. We thus establish a new state-of-the-art performance for large-scale deep reinforcement learning. Our results show the potential of curriculum learning in local learning settings with limited feedback and further bridges the gap between biologically plausible learning rules and error-backpropagation
NWO financiert drie projecten met betrokkenheid van Erasmus Universiteit Rotterdam en Erasmus MC
No full textSpike-based neuromorphic computing: An overview from bio-inspiration to hardware architectures and learning mechanisms
No full textThe endeavor to emulate the extraordinary efficiency and adaptability inherent in the human brain via spike-based neuromorphic computing presents significant potential across a diverse array of applications. The attainment of this objective necessitates the translation of biological principles into artificial systems, a task that continues to pose a complex challenge requiring a profound comprehension of the mechanisms by which neural systems produce robust computational outcomes. This tutorial paper provides a comprehensive overview of the foundational concepts and emerging design trends in spike-based neuromorphic computing, covering advances from materials and circuits to hardware architectures and learning mechanisms. It begins with an examination of key aspects of brain biology and their influence on neuromorphic design, followed by a brief discussion of biologically plausible neuron and synapse models. The paper then defines the core principles and defining attributes of neuromorphic computing, highlighting the trade-offs and design choices underlying current implementations. Building on these foundations, it explores the critical properties of neuromorphic systems, surveys a variety of learning algorithms, and reviews hardware-level realizations of bioinspired neurons and synapses. Subsequent sections discuss state-of-the-art spiking neural network architectures, mapping and compilation strategies, and representative application domains. By providing this end-to-end perspective, the article aims to guide the development of future neuromorphic systems that more closely emulate brain efficiency, scalability, and resilience
A critical analysis of deployed use cases for quantum key distribution and comparison with post-quantum cryptography
Get PDFQuantum Key Distribution (QKD) is currently being discussed as a technology to safeguard communication in a future where quantum computers compromise traditional public-key cryptosystems. In this paper, we conduct a comprehensive security evaluation of QKD-based solutions, focusing on real-world use cases sourced from academic literature and industry reports. We analyze these use cases, assess their security and identify the possible advantages of deploying QKD-based solutions. We further compare QKD-based solutions with Post-Quantum Cryptography (PQC), the alternative approach to achieving security when quantum computers compromise traditional public-key cryptosystems, evaluating their respective suitability for each scenario. Based on this comparative analysis, we critically discuss and comment on which use cases QKD is suited for, considering factors such as implementation complexity, scalability, and long-term security. Our findings contribute to a better understanding of the role QKD could play in future cryptographic infrastructures and offer guidance to decision-makers considering the deployment of QKD
Reading it like an open book: Single-trace blind side-channel attacks on garbled circuit frameworks
Get PDFGarbled circuits (GC) are a secure multiparty computation protocol that enables two parties to jointly compute a function using their private data without revealing it to each other. While garbled circuits are proven secure at the protocol level, implementations can still be vulnerable to side-channel attacks. Recently, side-channel analysis of GC implementations has garnered significant interest from researchers.We investigate popular open-source GC frameworks and discover that the AES encryption used in the garbling process follows a secret-dependent sequence. This vulnerability allows private inputs to be exposed through side-channel analysis. Based on this finding, we propose a side-channel attack on garbled circuits to recover the private inputs of both parties. Our attack does not require access to any plaintexts or ciphertexts in the protocol and is single-trace, adhering to the constraint that a garbled circuit can be executed only once. Furthermore, unlike existing attacks that can only target input non-XOR gates, our method applies to both input and internal non-XOR gates. Consequently, the secrets associated with every non-XOR gate are fully exposed as in an open book. Fundamentally, this work challenges the standard non-collusion assumption in multi-party computation, arguing for the necessity of extending it to the physical layer.We comprehensively evaluate our attack in various scenarios. First, we perform the attack on single-platform software implementations of standard AES and interleaved AES on a 32-bit ARM processor, achieving a 100% success rate in both cases. Next, we target a hardware implementation on a Xilinx Artix-7 FPGA, where the resolution of power consumption measurements and the number of samples are significantly limited. In this scenario, our attack achieves a success rate of 79.58%. Finally, we perform a cross-platform attack on two processors with different microarchitectures representing the two parties. The differing execution cycles and power sensors across the platforms increase the difficulty of side-channel analysis. Despite these challenges, our point-of-interest (POI) selection method allows our attack to achieve a 100% success rate in this scenario as well. We also discuss effective countermeasures that can be readily applied to GC frameworks to mitigate this vulnerability
Non-Interactive and Non-Destructive Zero-Knowledge proofs on quantum states and multi-party generation of authorized hidden GHZ states
Get PDFWe propose the first generalization of the famous Non-Interactive Zero-Knowledge (NIZK) proofs to quantum languages (NIZKoQS) and we provide a protocol to prove advanced properties on a received quantum state non-destructively and non-interactively (a single message being sent from the prover to the verifier). In our second orthogonal contribution, we improve the costly Remote State Preparation protocols [Cojocaru et al. 2019; Gheorghiu and Vidick 2019] that can classically fake a quantum channel (this is at the heart of our NIZKoQS protocol) by showing how to create a multi-qubit state from a single superposition. Finally, we generalize these results to a multi-party setting and prove that multiple parties can anonymously distribute a GHZ state in such a way that only participants knowing a secret credential can share this state, which could have applications to quantum anonymous transmission, quantum secret sharing, quantum onion routing and more
Inside every multithreaded program there are active objects struggling to get out
No full textMultithreading and actors offer different models of concurrency to the programmer. With multithreading, the programmer needs to deal with shared-state and data races, which make programs complex to understand, error-prone and challenging to verify, but potentially very efficient if these issues are mastered to perfection. On the other hand, actors—and their object-oriented incarnation as active objects,—which are inherently concurrent and protect their internal state against races, seem easy to understand and intuitive, but programs may be exposed to deadlocks due to callbacks. Is it possible to simply transition programs from the one concurrency model to the other at will, and thereby get the best of both worlds? We believe such a seamless transition between these concurrency models opens an interesting direction of research that remains to be investigated. As a step in this direction, this paper provides a high-level, informal outline of the translations between multithreading and active object concurrency, highlighting how intuitive or non-intuitive it is to move from one concurrency model to the other