A5069205069- Distributed systems and fault tolerance
- Age of Information Optimization
- Phonetics and Phonology Research
- Functional Brain Connectivity Studies
- Adversarial Robustness in Machine Learning
- Caching and Content Delivery
- French Urban and Social Studies
- Speech Recognition and Synthesis
- Mobile Agent-Based Network Management
- Covalent Organic Framework Applications
- Blockchain Technology Applications and Security
- Optimization and Search Problems
- Anomaly Detection Techniques and Applications
- IPv6, Mobility, Handover, Networks, Security
- Advanced Malware Detection Techniques
- Advanced Data Compression Techniques
- Wireless Communication Networks Research
- Speech and dialogue systems
- Cultural Insights and Digital Impacts
- Cognitive Functions and Memory
MGH Institute of Health Professions
2024
Google (United States)
2024
Institute of Science and Technology Austria
2023
École Polytechnique Fédérale de Lausanne
2023
École Nationale Supérieure des Arts Décoratifs
2021
This study examines the effectiveness of automatic speech recognition (ASR) for individuals with disorders, addressing gap in performance between read and conversational ASR. We analyze factors influencing this disparity effect mode-specific training on ASR accuracy.
Classical synchronous consensus algorithms are leaderless: processes exchange their proposals, retain the maximum value and decide when they see same choice across a couple of rounds. Indulgent more robust in that only require eventual synchrony, but however typically leader-based. Intuitively, this is weakness for slow leader can delay any decision. This paper asks whether, under it possible to deterministically solve without leader. The fact weakest failure detector one also eventually...
Consensus algorithms deployed in the crash fault tolerant setting chose a leader-based architecture order to achieve lowest latency possible. However, when wide area they face two key robustness challenges. First, lose liveness network is unreliable because rely on timeouts find leader. Second, cannot have high replication factor of load imposed leader-replica making it bottleneck. This effectively limits allowed, for given level throughput, thus lowering tolerance threshold. In this paper,...
Download This Paper Open PDF in Browser Add to My Library Share: Permalink Using these links will ensure access this page indefinitely Copy URL DOI
State-of-the-art deep learning models for tabular data have recently achieved acceptable performance to be deployed in industrial settings. However, the robustness of these remains scarcely explored. Contrary computer vision, there is date no realistic protocol properly evaluate adversarial due intrinsic properties such as categorical features, immutability, and feature relationship constraints. To fill this gap, we propose CAA, first efficient evasion attack constrained models. CAA an...
Sharding distributed ledgers is a promising on-chain solution for scaling blockchains but lacks formal grounds, nurturing skepticism on whether such complex systems can scale securely. We fill this gap by introducing the first framework as well roadmap to robust sharding. In particular, we define properties sharded should fulfill. build upon and extend Bitcoin backbone protocol defining consistency scalability. Consistency encompasses need atomic execution of cross-shard transactions...
Consensus algorithms are deployed in the wide area to achieve high availability for geographically replicated applications. Wide-area consensus is challenging due two main reasons: (1) low throughput latency overhead of client request dissemination and (2) network asynchrony that causes protocols lose liveness. In this paper, we propose Mandator Sporades, a modular state machine replication algorithm enables performance resiliency wide-area setting. To address challenge, Mandator, novel...