Recursive Binding on a Budget: Subspace Carving in Order-p Tensor Memories

In three linesOSC (Orthogonal Subspace Carving) resolves the trade-off between Tensor Product Representations (exponential growth) and Vector Symbolic Architectures (noisy compression). By projecting onto the null space of the role basis before aggregating into fixed order-p tensors, OSC decouples tensor order from structural depth, enabling deep recursive binding with constant memory footprint.

Summary generated by Claude — human-verified