The Spacetime Cognition Deficit: Why Architecture Matters for General Embodied Intelligence

Fuente: Zenodo
Gespeichert in:
Bibliographische Detailangaben
1. Verfasser: Li, Hengbo
Format: Recurso digital
Veröffentlicht: Zenodo 2026
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866901998812528640
author Li, Hengbo
author_facet Li, Hengbo
contents We propose the spacetime cognition deficit hypothesis: conventional Transformer architectures are structurally limited in achieving general embodied intelligence because they lack intrinsic spacetime modeling capability, and this limitation is architectural rather than parametric. We present three lines of evidence from the LSH-SFA series and outline a concrete validation path using the LSH project's existing 3D world engine with pluggable physics backends, proposing three experiments that can definitively test the hypothesis. LSH-SFA, with its intrinsic spacetime encoding, PDE-driven evolution, light-cone causality, and energy conservation, provides the general-purpose architectural foundation that embodied intelligence requires.
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_20129066
institution Zenodo
language
publishDate 2026
publisher Zenodo
record_format zenodo
spellingShingle The Spacetime Cognition Deficit: Why Architecture Matters for General Embodied Intelligence
Li, Hengbo
spacetime cognition deficit
general embodied intelligence
Transformer limitations
physical priors
wave packet representation
LSH-SFA
architectural diversity
physical simulation
embodied cognition
general-purpose foundation
We propose the spacetime cognition deficit hypothesis: conventional Transformer architectures are structurally limited in achieving general embodied intelligence because they lack intrinsic spacetime modeling capability, and this limitation is architectural rather than parametric. We present three lines of evidence from the LSH-SFA series and outline a concrete validation path using the LSH project's existing 3D world engine with pluggable physics backends, proposing three experiments that can definitively test the hypothesis. LSH-SFA, with its intrinsic spacetime encoding, PDE-driven evolution, light-cone causality, and energy conservation, provides the general-purpose architectural foundation that embodied intelligence requires.
title The Spacetime Cognition Deficit: Why Architecture Matters for General Embodied Intelligence
topic spacetime cognition deficit
general embodied intelligence
Transformer limitations
physical priors
wave packet representation
LSH-SFA
architectural diversity
physical simulation
embodied cognition
general-purpose foundation
url https://doi.org/10.5281/zenodo.20129066