- Rating
- 2
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 1
Read review
Paper summary
SpecSIA scores each client by the inner product between the normalized target gradient and the client's normalized current-round update (RCA, i.e., cosine similarity), after applying a median-calibrated spectral shrinkage to the normalized updates (ASF). The target is attributed to the highest-scoring client. Averaged over each of three datasets, it reports higher ASR and lower Temporal Std than loss-based SIA.
Critical strengths
The method is simple and cheap (one target gradient plus a solve), and uses the same server view as loss-based SIA, with no shadow models or historical updates. Evaluating SIA beyond mean ASR, in terms of temporal stability and cross-configuration consistency, is a meaningful perspective. RCA (and hence SpecSIA) yields clear gains over loss-based SIA on MNIST and Synthetic.
Critical weaknesses
- Flawed motivation. Since is common to all clients, ranking by the exact round-anchored contribution gives , i.e., loss-based SIA. The gain thus stems from linearization or normalization, not from isolating "the contribution introduced by the current client update"; indeed, in Table 1 the gain on CIFAR-10 and Synthetic is near zero at (where linearization is most accurate) and largest at . Ablations and are missing.
- ASF brings no meaningful benefit. ASF changes ASR by and Temporal Std by , far below RCA's / , without significance tests. ASR rises monotonically with weaker filtering (larger ); the authors concede dominant directions "may still contain useful source-discriminative information", contradicting ASF's premise. The Rank IQR gain (2.00→1.00) rests on coarse integer ranks; per dataset (Table 5), ASF worsens Mean Rank on all three datasets and Rank IQR on CIFAR-10. Yet ASF is the title component.
- Limited novelty. RCA is gradient–update cosine similarity (cf. TracIn / influence functions, not discussed). The ASF score equals, up to a positive constant, the ridge-regression coefficients of on .
- Theory does not support the method. The error bound applies to the unnormalized score, not the normalized score actually used. The scale-invariance result is vacuous because when .
- Limited experiments. The evaluation covers only FedAvg with , full participation and small models. Additional experiments under broader FL settings are needed.
Questions for authors
Please refer to weakness.