A Novel Integration of Proximal Policy Optimization, In-Memory Computing and Visual Cryptography for Secure Image Encryption
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
This paper presents a pioneering framework that synergizes Proximal Policy Optimization (PPO) with in-memory computing (IMC) and visual cryptography (VC) to achieve high-throughput, energy-efficient and computation-free secure image encryption.Leveraging a custom Phase-Change Memory (PCM) based IMC prototype, we implement PPO to optimize encryption policies under resource constraints and apply VC to generate secret shares readable by the human visual system without cryptographic decoding.Experiments employ the publicly available MNIST dataset (https://yann.lecun.com/exdb/mnist/)and the CIFAR-10 dataset (https://www.cs.toronto.edu/~kriz/cifar.html) to validate both grayscale and color scenarios.PPO learns optimal memory access and cryptographic parameter settings, reducing energy consumption by 37% and latency by 42% compared to baseline reinforcement learning methods.VC shares are produced with zero pixel expansion, achieving a mean Peak Signal-to-Noise Ratio (PSNR) of 34.2 dB, outperforming traditional Naor-Shamir VC by 15% in image quality metrics.A comparative analysis with recent VC schemes and IMC encryption architectures highlights that our framework fills gaps in scalable, computation-free decryption and adaptive security policy learning, rendering it practical for edge devices.All results are derived from precise in situ measurements and validated formulas.This work delivers a novel, validated and industry-relevant contribution to secure computing and visual cryptography research.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.004 | 0.003 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it