REVIEW 6 cited by
Complete Dictionary Learning via $\ell_p$-norm Maximization
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
abstract
Dictionary learning is a classic representation learning method that has been widely applied in signal processing and data analytics. In this paper, we investigate a family of $\ell_p$-norm ($p>2,p \in \mathbb{N}$) maximization approaches for the complete dictionary learning problem from theoretical and algorithmic aspects. Specifically, we prove that the global maximizers of these formulations are very close to the true dictionary with high probability, even when Gaussian noise is present. Based on the generalized power method (GPM), an efficient algorithm is then developed for the $\ell_p$-based formulations. We further show the efficacy of the developed algorithm: for the population GPM algorithm over the sphere constraint, it first quickly enters the neighborhood of a global maximizer, and then converges linearly in this region. Extensive experiments will demonstrate that the $\ell_p$-based approaches enjoy a higher computational efficiency and better robustness than conventional approaches and $p=3$ performs the best.
Forward citations
Cited by 6 Pith papers
-
Fixed Budget vs. Covering Target: The Partial Set Cover Boundary for Bounded VC-Dimension
Partial Set Cover is inapproximable below factor 2 even at VC-dimension 7, while bounded semi-ladder index restores a k+1-sets covering target and yields new EPAS results.
-
On the Computational Complexity of (Extended) Threshold Dimension and (Semi-)Ladder Index
Threshold dimension is NP-complete; Extended threshold dimension is NP-hard and co-NP-hard; both, plus ladder/semi-ladder indices, resist |X|^o(1) and FPT o(k) approximation under Gap-ETH.
-
Tarallo: Evading Behavioral Malware Detectors in the Problem Space
Tarallo evades RNN-based behavioral malware detectors by injecting API calls selected with a new position-sensitive gradient attack, reaching up to 99% evasion in feature and problem space tests.
-
Does Runtime Topology Context Improve LLM-Generated Kubernetes Security Patches?
Providing LLMs with live cluster topology and service-account context raised correct Kubernetes security patches from 11.1% to 78.0% on topology-dependent findings, with no change on a topology-independent control.
-
The Vulnerability With No CVE: Managing Persistent Gaps Between Mandate and Authority in AI Coding Agents
The paper defines a new vulnerability-management object, the agentic posture vulnerability, with a thresholded definition, lifecycle, record schema, and testable research agenda for AI coding agents.
-
PECR: A Reproducible Specification and Synthetic Stress Test of Telemetry-Informed Vulnerability Prioritization for SD-WAN
PECR is a transparent nine-factor vulnerability prioritization protocol for SD-WAN, with synthetic stress tests showing it ranks differently than CVSS/EPSS/KEV queues, but no real outcome validation.
Discussion (0). Continue with ORCID to comment.