Research map / Optimization / Implicit bias and generalization
Information-Theoretic Compressive Measurement Design
Liming Wang, Minhua Chen, Miguel R. D. Rodrigues, David S. Wilcox, Robert Calderbank, Lawrence Carin
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use ITM with b=0.5 instead of pure signal recovery design | pure signal recovery (PSR) · Satellite classification, L=5, m=3 | Classification accuracy: 0:8400 -> 0:8575 | Improved |
| use ITM with b=0.5 instead of pure classification design | pure classification (PC) · Satellite signal recovery, L=5, m=3 | Fractional error: 0:5233 -> 0:4897 | Improved |
| ITM b=0.5 versus pure classification design in accuracy | pure classification (PC) · Satellite classification, L=5, m=3 | Classification accuracy: 0:8600 -> 0:8575 | Hurt |
| ITM b=0.5 versus LDA for classification | LDA · Satellite classification, L=5, m=3 | Classification accuracy: 0:8535 -> 0:8575 | Improved |
| ITM b=0.5 versus LDA for recovery | LDA · Satellite signal recovery, L=5, m=3 | Fractional error: 0:6876 -> 0:4897 | Improved |
| increase number of measurements m from 1 to 2 for ITM b=0.5 (ablation) | ITM · Satellite classification, L=5 | Classification accuracy: 0:4935 -> 0:8265 | Improved |
| raise b from 1 to 100 (lower rank) (ablation) | ITM · Satellite classification vs rank | Classification accuracy: 0.9012 -> 0.6343 | Hurt |
| rank-reduce projection matrix after design (ablation) | ITM b=10 · Satellite classification | Classification accuracy: 0.8642 -> 0.8622 | No clear effect |
| raise b from 50 to 500 on USPS (lower rank) (ablation) | ITM · USPS classification vs rank | Classification accuracy: 0.9131 -> 0.3416 | Hurt |
| rank-reduce projection matrix on USPS (ablation) | ITM b=50 · USPS classification | Classification accuracy: 0.9131 -> 0.9119 | No clear effect |
| use LDA and IDA instead of ITM for USPS recovery | LDA, IDA · USPS signal recovery | Fractional error | Hurt |
| energy constraint L=1 instead of L=5 (ablation) | ITM · Satellite | Classification accuracy | No clear effect |
Closest papers by meaning
- The Price of Sparsity: Sufficient Conditions for Sparse Recovery using Sparse and Sparsified Measurements (NeurIPS 2025 · similarity 0.84)
- Universality in Learning from Linear Measurements (NeurIPS 2019 · similarity 0.83)
- Two New Approaches to Compressed Sensing Exhibiting Both Robust Sparse Recovery and the Grouping Effect (JMLR 2017 · similarity 0.83)
- A Statistical Experimental Design Method for Constructing Deterministic Sensing Matrices for Compressed Sensing (JMLR 2024 · similarity 0.83)
- Exact Recovery of Sparse Binary Vectors from Generalized Linear Measurements (ICML 2025 · similarity 0.82)
- Information-Driven Design of Imaging Systems (NeurIPS 2025 · similarity 0.81)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
