- A feedback-driven adaptive learning binary-classifier that based on error input autonomously alters its decision boundary by implementing Max-Initialized Decremental Search (MIDS) and resets the control loop for repeated adaptive cycles.
🤖 Adaptive learning system implementing MIDS algorithm
The comparator output exhibited a static hazard due to unequal propagation delays on the threshold update signals. Since the edge-detection circuitry monitored transitions continuously, the transient pulse was interpreted as a legitimate boundary crossing. A hazard detection mechanism was therefore added to invalidate simultaneous activation of both edge detectors, preventing false convergence.
- Stage 0 (v0.x): Strict Boolean pattern relation analyzer. No learning, no noise tolerance, decision boundaries fixed by structural wiring.
- Stage 1 (v1.0): Popcount based similarity and a variable threshold to alter the decision boundary. Introduces noise tolerance and ability to change the decision output without structural changes.
- Stage 2 (v1.1): Cybernetic Feedback-driven adaptive learning. System alters its decision boundary based on external feedback to correct its decision output.
-
Version 0: A pattern relation analyzer that classifies how an input pattern relates to a stored pattern, enforces rule based recognition rather than learning.- Detector_v0.0 -> Recognizes the exact pattern and sub-patterns if they are inside the boundary set up by weights-grid.
- Detector v0.1 -> Recognizes the exact pattern and super-patterns if they are outside the boundary set up by weights-grid.
- Detector v0.2 -> Classifies the input as a sub-pattern, super-pattern, anti-pattern or equivalence precisely through a 2-POV logical analysis.
-
Version 1: Pop-count based judgement against a variable Threshold instead of perfect equivalence check and cybernetic feedback-driven adaptive learning.- Detector_v1.0 -> Recognizes the pattern if total number of matched pixels are greater than the set threshold which can vary giving us ability to control the decision output.
- Detector_v1.1 -> A feedback-driven adaptive system that autonomously adjusts its decision boundary to correct its output, using algorithms optimized for hardware constraints.
🧩 Block Diagram - Detector_v1.0 (Manually Alterable Decision Boundary)
| Property | MIDS | SATU |
|---|---|---|
| Correction Speed | O(N) | O(1) |
| State Awareness | None | Current & desired output |
| Direction | Always starts from max, decrements | Sets to M or M-1 as needed |
| Initialization Bias | Instant correction for false positives | None |
| Hardware Complexity | Low - decrementer only | Higher - decrementer + decision logic |
| Guaranteed Convergence | Yes | Yes |
⏱️ Correction Speed Complexity Comparison
Equivalence & Sub-Pattern Recognition 🔹
Equivalence & Super-Pattern Recognition 🟦
Equivalence & Super & Sub & Anti-Pattern Recognition 🔹🟦
Manually Alterable Decision Boundary ⚖️
To verify hardware realizability, all detector variants were synthesized using Yosys, technology mapped to the Sky130HD standard-cell library, and analyzed using static timing and power estimation. The resulting gate-level netlists were used to compare architectural complexity, silicon area, timing characteristics, power consumption, and estimated operating frequency across the evolution of the Cybernetic Classifier.
Technology: Sky130HD
| Version | Module | Cells | Area | Critical Path Delay | Power | Key Hardware Structures |
|---|---|---|---|---|---|---|
| Detector v0.0 | Eq/Sub Recognizer | 8 | 127.6224 µm² | 0.46 ns | 30.8 µW | 3 AND, 2 NOT, OR, Reduction-AND |
| Detector v0.1 | Eq/Super Recognizer | 8 | 127.6224 µm² | 0.46 ns | 30.8 µW | 3 AND, 2 NOT, OR, Reduction-AND |
| Detector v0.2 | Multi-POV Classifier | 28 | 230.2208 µm² | 1.16 ns | 60.6 µW | Dual Recognition Engines, Decision Logic, 11 MUXes |
| Detector v1.0 | Pop-Count Recognition | 22 | 970.9312 µm² | 4.58 ns | 599 µW | 15 Adders, Comparator, Threshold Logic |
| Category | Result |
|---|---|
| Smallest Design | Eq/Sub Recognizer (127.62 µm²n) |
| Smallest Design | Eq/Super Recognizer (127.62 µm²) |
| Largest Design | Pop-Count Recognition (970.93 µm²) |
| Fastest Design | Eq/Sub Recognizer (Fmax ≈ 1/0.46 ns ≈ 2.17 GHz) |
| Fastest Design | Eq/Super Recognizer (Fmax ≈ 1/0.46 ns ≈ 2.17 GHz) |
| Slowest Design | Pop-Count Recognition (Fmax ≈ 1/4.58 ns ≈ 218 MHz) |
| Lowest Power | Eq/Sub Recognizer (30.8 µW) |
| Lowest Power | Eq/Super Recognizer (30.8 µW) |
| Highest Power | Pop-Count Recognition (599 µW) |
| Most Arithmetic-Heavy | Pop-Count Recognition (15 Adders, Comparator, Threshold Logic) |
| Most Decision-Heavy | Multi-POV Classifier (11 MUXes) |
| Largest Cell Count | Multi-POV Classifier (28 Cells) |
Detector_v1.0 Synthesized - Yosys ✅
- Stage 2 in progress - MIDS: Developed ✓ | SATU: In Development
Download → download_repos.bat
Double-click it and pick the repo(s) you want.
Download → download_repos.sh
bash
chmod +x download_repos.sh
./download_repos.sh
Always downloads the latest version.
- Source code and HDL files are licensed under the MIT License.
- Documentation, diagrams, images, and PDFs are licensed under Creative Commons Attribution 4.0 (CC BY 4.0).