RF-Rate Hybrid CNN Accelerator Based on Analog-CMOS and Xilinx RFSoC
Bibliographic record
Abstract
The superior performance of deep learning (DL) has sent shock waves in the machine learning community. The high adoption rate of DL has set new demands on computational throughput, latency, and power efficiency of the computing infrastructure. In addition to conventional approaches to acceleration of the inference component of DL systems based on GPUs, cloud computing, ASIC/FPGAs and custom vector processors (such as tensor processing units), there is renewed interest in high-frequency analog circuits for DL inference. Analog computing is a potential candidate for meeting challenging requirements in throughput, latency and power efficiency. Because DL inference has superior noise resilience and relatively low accuracy needs (typically less than 8 bits), analog circuits can provide a promising alternative to all-digital accelerators. This paper presents early work on the design of an analog CMOS accelerator that performs analog convolution and decision operations in parallel and in real-time by pairing a high-frequency operational amplifier-based CNN filtering kernel with a rectified linear unit (ReLu) non-linearity based on an active precision rectifier circuit. The analog accelerator was designed in a 45 nm CMOS process and simulated in Cadence Spectre. Image convolution results are presented and compared with MATLAB simulations. The proposed solution also employs Xilinx RF System-on-Chip (SoC) devices based on the Xilinx ZCU1285 RFSoC platform to interface digital inputs and outputs with the proposed RF-rate analog inference accelerator.
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.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.015 | 0.002 |
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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".