Maya-Śūnyatā: Karma-Weighted Synaptic Pruning for Class-Incremental Learning in Affective Spiking Neural Networks

Version 2 (2026-08-04). Revised to the NLL Universal Paper Format v6. The original text is retained in full; nothing has been deleted. Corrections appear as marked blocks placed at the section that carries the claim, and each one states what the paper said, what the data show, the corrected claim, why it happened, and what still stands. This version withdraws the framing that made a negative result unfalsifiable.Every pruning condition is 5–9 pp worse than not pruning, and the reason is now given: Σ|Δw| is the Synaptic Intelligence path integral, which SI uses to protect high-movement weights and this paper uses to delete them; the stated explanation for condition C ≡ A is also arithmetically wrong and the real cause is given from the code.An unresolved conflict is flagged rather than silently reconciled: this paper reports condition D at 9.73% / 62.01% pruned, while P09 reports the same configuration at 12.46% / 85.19%. Everything below this line is the original description from version 1. It is retained unchanged for the record. Where it conflicts with the corrections above, the corrections stand. Two framings it repeats have since been withdrawn in full — the Bhaya Quiescence Law and the Buddhi S-Curve. Both are addressed in Maya-Meta P1, now superseded, and in the self-audit of Maya-Meta P2. Maya-Śūnyatā introduces a structured synaptic pruning mechanism for spiking neural networks that must learn new categories of information over time without forgetting what they have already learned — a fundamental challenge in artificial intelligence called class-incremental learning. The architecture adds two new computational primitives: Karma, a second-order signal that tracks how much each synaptic connection has been pulled in conflicting directions across successive learning tasks, and Śūnyatā, a pruning mechanism that zeros out synapses whose accumulated interference exceeds a threshold modulated by the network's developmental maturity, producing emergent behavior in which younger network states prune aggressively and older ones conservatively — a dynamic grounded biologically in microglial phagocytosis via the C1q/C3 complement cascade. A seven-condition ablation study on Split-CIFAR-100 (10 tasks) reveals that Karma must accumulate across a full task boundary to reach pruning threshold, as continuous batch-level pruning never fires and exactly reproduces the zero-pruning baseline at 15.19% average accuracy, while the full Śūnyatā mechanism with Karma accumulation achieves measurable improvements over this baseline. This is the eighth paper in the Maya Research Series, a thirteen-paper program from Nexus Learning Labs implementing the Advaita Vedantic Antahkarana as computational primitives in spiking neural networks, with the Bhaya Quiescence Law (β* ≤ 0.32%) and Buddhi S-Curve Determinism (R² = 1.0000) confirmed across all series entries to date. The series advances a principled research agenda for deriving novel neural network mechanisms from classical Indian philosophical frameworks as a source of structured biological and computational inspiration. Series: Part of the Maya Research Series — 13 papers implementing the Advaita Vedantic Antahkarana as computational primitives in spiking neural networks. Bhaya Quiescence Law (β* ≤ 0.32%) and Buddhi S-Curve Determinism (R²=1.0000) confirmed across all papers. Links: GitHub Repository (private — to request access: email research@nexuslearninglabs.in with subject Code Access Request — Maya-Shunyata and your research context) | Interactive Dashboard | FAQ | Full Series Index — venky2099.github.io Nexus Learning Labs, Bengaluru · UDYAM-KR-02-0122422 · BHASKAR IN-0526-9452JSORCID: 0000-0002-3315-7907 · VAIRAGYA_DECAY_RATE = 0.002315 — an ORCID-derived provenance mark, not an experimental parameter. Each paper's disclosure block states whether it reached a result in that paper.Canary: MayaNexusVS2026NLL_Bengaluru