Applying deep compression to enable spoken language edge inference
Bibliographic record
Abstract
memory compression and inference performance at the pruning stage provides more than 1.5x compression for a 1% increase in word error rate.After applying this optimization pipeline to the modified Deep Speech 2 model, its memory consumption is on the order of most modern cache sizes.Thus, rendering it deployable as an embedded offline voice recognizer.v I would like to take this opportunity to acknowledge friends, family, and professors alike for their unwavering support during the trying pandemic-filled past two years.Their emotional support, countless acts of motivation, and research discussions have helped me craft the kind of research inquiry presented in the latter sections of this work.I would like to acknowledge my thesis supervisor, Prof. Zeljko Zilic, for his constant guidance and discussions on interesting research topic formulation and execution.I would also like to acknowledge Prof. Brett Meyer for providing research background resources related to data presentation, data collection, and general research domain and angle of attack.From my research lab to other joint labs under different professors, I would like to acknowledge my friend/lab-mate, Hamza Mian, for suggesting the kind of idea applied/tested in this research work on a domain that is completely different than what that idea was originally intended for.In addition, I would also like to acknowledge my Ph.D. senior, Shibam Debbarma, for his constant
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 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 | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| 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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, 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".