- Rating
- 2
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 1
Read review
Paper summary
SkewSpike combines a skewed-tree dictionary, partial pattern matching, stochastic residual pruning, and threshold calibration to compress inter-layer spike traffic. Experiments evaluate compression accuracy trade-offs across several SNN architectures and datasets, alongside synthesized-router estimates of dynamic communication energy.
Critical strengths
- The paper directly targets inter-layer AER traffic, addressing a practical bottleneck beyond weight compression.
- The evaluation covers fully connected and convolutional SNNs on static-image and neuromorphic datasets.
- Component ablations and parameter sweeps expose important compression–accuracy trade-offs.
Critical weaknesses
-
The autotuner uses task accuracy on 50% of the evaluation data, but final results are not clearly restricted to an independent, untouched test split.
-
Dictionary replacement alone yields only 0.03% compression in Table 3; the skewed tree’s benefit over random pruning with the same threshold calibration remains unisolated.
-
Compression accounting mixes event counts and bit lengths, while energy estimates omit codec and dictionary costs; net end-to-end efficiency gains remain unestablished.
-
DVS128-Gesture accuracy drops by 34.73 percentage points at SS50, without deep-model uncertainty estimates; VGG11 compression also differs between Table 2 and Table 4’s counts.
-
Appendix A.3 attributes FP-Max to an FP-Growth reference, while Appendix A.2 cites Werbos’s BPTT paper for the specific ATan surrogate, leaving these algorithm choices inaccurately supported.
Questions for authors
-
Which samples and labels are used for dictionary construction, autotuning, and final evaluation, and can all methods be evaluated on the same untouched test split?
-
Can you compare against random pruning with the same threshold calibration and traffic budget, and against balanced-tree or flat dictionaries, to isolate the skewed-tree contribution?
-
What net latency and energy gains remain under common bit-level accounting that includes codec and dictionary costs? Does FSPC’s timing include pattern mining performed offline by SkewSpike?
-
Why does VGG11/CIFAR-10 SS50 report 50.12% compression in Table 2 but approximately 54.59% from Table 4’s counts, and how stable are the deep-model results across seeds?
-
How are empty transactions and p=0 handled, and how many spurious spikes does partial matching introduce? What denominator is used for the constant compression ratio in Figure 2(d)?
We thank the reviewer for their critical review. We have addressed the comments with new no-dictionary baselines, held-out split validations, alternative dictionary benchmarks, multi-seed stochastic tests, and citation fixes in our revised manuscript PDF.
Final performance is evaluated across the full test set and validated on the strictly untouched 50% held-out split (Appendix A.9, Table 11, Page 20):
A. No-Dictionary Baseline at Equal 50% Budget (Table 3, Page 9):
Takeaway: Calibration prevents collapse (+34.06 pp VGG16, +7.14 pp ResNet18), while the dictionary offers a modest, consistent accuracy margin (+0.22 to +0.66 pp in CNNs) and +2.94% extra structural compression on MNIST.
B. Alternative Dictionaries on SEW-ResNet18 (Appendix A.8, Table 10, Page 20):
Reconciliation of Table 2 vs. Table 4 (Weakness 4, Q4): Table 2 reports per-batch average compression (50.12%), whereas Table 4 reports cumulative dataset-wide counts from seed 42 (109.96M -> 49.93M spikes, 54.59% reduction) from the standalone hardware-trace export for QuestaSim/PrimeTime simulation (clarified on Page 9).
Bibliographic Corrections (Weakness 5): (1) FP-Max now cites Grahne & Zhu (2003) (Page 11, Line 583; Page 15, Line 757). (2) ATan surrogate cites Fang et al. (2021) (Page 11, Line 569; Page 13, Line 696), retaining Werbos (1990) solely for BPTT.
Bit-Level Accounting & Codec Overhead (Weakness 3, Q3): AER spikes and symbols are both 16-bit flits (Table 8), making event reduction equal to bit reduction. For N=10,000 and |V|=5 symbols (Table 10), dictionary SRAM requires only 6.10 KB (<156.25 KB even at |V|=128). FSPC latency (2.38-17.70s/batch) is pure online CPU inference; SkewSpike executes in 2.36-2.79 ms on GPU.