In this paper we study the training dynamics for gradient flow on over-parametrized tensor decomposition problems. Empirically, such training process often first fits larger components and then discovers smaller components, which is similar to a tensor deflation process that is commonly used in tensor decomposition algorithms. We prove that for orthogonally decomposable tensor, a slightly modified version of gradient flow would follow a tensor deflation process and recover all the tensor components.
In the first part we cover five current specific challenges through examples: (1) discrimination (e.g., facial recognition, justice, sharing economy, language models); (2) phrenology (e.g., biometric based predictions); (3) unfair digital commerce (e.g., exposure and popularity bias); (4) stupid models (e.g., Signal, minimal adversarial AI) and (5) indiscriminated use of computing resources (e.g., large language models). These examples do have a personal bias but set the context for the second part where we address four generic challenges: (1) too many principles (e.g., principles vs.
CS Department members, please join us for this special opportunity to come together in fellowship and welcome our newest faculty, graduate students, undergraduate students, postdocs, and staff. The meeting will begin at 12:00 pm, with remarks from the department chair and presentation of awards. Boxed lunches will be provided for you to take with you and enjoy.
Fueled by massive amounts of data, models produced by machine-learning (ML) algorithms, especially deep neural networks (DNNs), are being used in diverse domains where trustworthiness is a concern, including automotive systems, finance, healthcare, natural language processing, and malware detection. Of particular concern is the use of ML algorithms in cyber-physical systems (CPS), such as self-driving cars and aviation, where an adversary can cause serious consequences. Interest in this area of research has simply
Structured encryption (STE) schemes encrypt data structures in such a way that they can be privately queried. Special cases of STE include searchable symmetric encryption (SSE) and graph encryption. Like all sub-linear encrypted search solutions, STE leaks information about queries against persistent adversaries. To address this, a line of work on leakage suppression was recently initiated that focuses on techniques to mitigate the leakage of STE schemes. A notable example is the query equality suppression framework (Kamara et al.