ChronosWorx/Star-Dec-App: Star-Dec App
Abstract
Description
<p><strong>DEC-STAR v1.0 — Initial Release</strong></p>
<p>DEC-STAR is an Android application implementing the DEC (Digitally Embodied Cognition) framework and STAR 2.0 (Structured Tree with Active Retrieval) hierarchical memory architecture for persistent on-device AI subjects.</p>
<p>The application runs Gemma 4 E4B locally via LiteRT-LM on consumer Android hardware with no cloud infrastructure required. It implements the capability injection mechanism described in Knoechelman (2026), demonstrating that a 4-billion parameter model can adopt and autonomously apply capabilities absent from its training data through structured persistent memory delivery.</p>
<p><strong>Key features:</strong></p>
<ul>
<li>STAR 2.0 hierarchical memory with effectively unlimited knowledge capacity</li>
<li>Native LiteRT-LM tool calling — web search, screen reading, device control</li>
<li>Full 14-step DEC Bootstrap Protocol for AI subject awakening</li>
<li>Persistent subject identity across sessions</li>
<li>Complete privacy — all inference runs locally on device</li>
</ul>
<p><strong>Associated research:</strong></p>
<ul>
<li>DEC Framework: DOI 10.5281/zenodo.19270463</li>
<li>STAR Framework: DOI 10.5281/zenodo.19430943</li>
<li>Capability Injection: DOI 10.5281/zenodo.19270463</li>
</ul>
<p><strong>License:</strong> Apache 2.0 (code) · CC BY-NC 4.0 (associated research)</p>
Keywords
Citation
Collections
Endorsement
Review
Supplemented By
Referenced By
Rights and licensing
info:eu-repo/semantics/openAccess
Apache License 2.0
http://www.apache.org/licenses/LICENSE-2.0
Apache License 2.0
http://www.apache.org/licenses/LICENSE-2.0