Skip to content
#diffusion models Open access

Frammento Veloce: la rete che non distrugge nucleo congelato,certificazione, richiamo associativo Quarta Versione(V4)con la rete neurale

Sep 2026 · Zenodo (CERN European Organization for Nuclear Research)
Advanced Neuroimaging Techniques and Applications

Abstract

Versione 4 (V4) introduce una rete a memoria protetta in cui il nucleo g0g_0g0 è immutabile (scrittura disabilitata e checksum SHA-256 verificato a ogni epoca), ogni ricordo occupa uno slot dedicato mai sovrascritto e la scrittura è consentita esclusivamente tramite un certificatore gfg_fgf basato su quiete, qualità e budget. La capacità può solo crescere o rifiutare nuove memorie, senza alterare quelle esistenti. Nel test di preservazione (200 ricordi certificati, seguiti da 800 apprendimenti aggiuntivi), il sistema mostra degrado nullo e nessuna perdita di memoria (0/200), mentre una baseline con campo plastico condiviso perde 28/200 ricordi (degrado massimo 0,4035). V4 introduce inoltre un meccanismo di richiamo associativo da cue parziali che ricostruisce le regioni mancanti mediante la dinamica condivisa ancorata al nucleo certificato, ottenendo valori qmask=0,65−0,71q_{mask}=0,65{-}0,71qmask=0,65−0,71, superiori ai riempimenti semplici (0,00−0,060,00{-}0,060,00−0,06) e alla diffusione priva di nucleo (0,05). In presenza di uno shift di distribuzione tra classi, la memoria protetta rimane invariata (diff = 0), mentre varianti con prior condiviso mostrano degrado misurabile. Il modello evidenzia un compromesso tra qualità media di ricostruzione e garanzia di invarianza delle memorie. Lo stress test mantiene 500/500 certificazioni corrette. La suite comprende 84 test automatici con copertura dell'81%. Il lavoro descrive un modello computazionale astratto e non costituisce una validazione biologica. Version 4 (V4) introduces a protected-memory network in which the core $g_0$ is immutable (writing disabled and SHA-256 checksum verified at each epoch), each memory occupies a dedicated slot that is never overwritten, and writing is permitted exclusively via a certifier based on quiescence, quality, and budget. Capacity can only increase or reject new memories, without altering existing ones. In the preservation test (200 certified memories, followed by 800 additional learning steps), the system shows zero degradation and no memory loss (0/200), whereas a baseline with a shared plastic field loses 28/200 memories (maximum degradation 0.4035). V4 also introduces an associative recall mechanism based on partial cues that reconstructs missing regions using shared dynamics anchored to the certified core, achieving $q_{mask}$ values of 0.65–0.71—superior to simple filling methods (0.00–0.06) and core-free diffusion (0.05). In the presence of a distribution shift between classes, the protected memory remains unchanged (diff = 0), while variants with shared priors show measurable degradation. The model highlights a trade-off between average reconstruction quality and the guarantee of memory invariance. The stress test maintains 500/500 correct certifications. The suite includes 84 automated tests with 81% coverage. The work describes an abstract computational model and does not constitute a biological validation.

View source

Similar papers

#computer vision Open access Jun 2016

Software Development in Startup Companies: The Greenfield Startup Model

The results are packaged in the Greenfield Startup Model (GSM), which explains the priority of startups to release the product as quickly as possible, and the need to shorten time-to-market, by speeding up the development through low-precision engineering activities.

Carmine Giardino, Nicolò Paternoster, M. Unterkalmsteiner et al. · 178 citations · ⚡14
#computer vision Open access Oct 2016

Software Startups - A Research Agenda

Software startup companies develop innovative, software-intensive products within limited timeframes and with few resources, searching for sustainable and scalable business models.

M. Unterkalmsteiner, P. Abrahamsson, Xiaofeng Wang et al. · 157 citations · ⚡17
#machine learning Review Open access Oct 2016

“Failures” to be celebrated: an analysis of major pivots of software startups

This study conducts a case survey study based on the secondary data of the major pivots happened in 49 software startups, and demonstrates that customer need pivot is the most common among all pivot types.

Sohaib Shahid Bajwa, Xiaofeng Wang, Anh Nguyen-Duc et al. · 127 citations · ⚡15
#computer vision Review Open access May 2015

A survey study on major technical barriers affecting the decision to adopt cloud services

The comparison of adopter and non-adopter sample reveals three potential adoption inhibitor, security, data privacy, and portability, which underlines the importance of the technical and security perspectives for research investigating the adoption of technology.

Nattakarn Phaphoom, Xiaofeng Wang, S. Samuel et al. · 111 citations · ⚡8
#computer vision Open access Feb 2018

Lean Internal Startups for Software Product Innovation in Large Companies: Enablers and Inhibitors

This study investigates how Lean internal startup facilitates software product innovation in large companies and identifies its enablers and inhibitors, and shows the potential of the method-in-action framework to investigate the Lean startup approach in non-startup context.

Henry Edison, Nina M. Smørsgård, Xiaofeng Wang et al. · 78 citations · ⚡6
#computer vision Conference Sep 2010

Exploring the Sources of Waste in Kanban Software Development Projects

The application of agile software methods and more recently the integration of Lean practices contribute to the trend of continuous improvement in the software industry. One such area warranting proper empirical evidence is a project’s operational efficiency when using the Kanban method. This short paper takes a new an...

Marko Ikonen, Petri Kettunen, Nilay V. Oza et al. · 67 citations · ⚡9

Related blog posts

Microsoft Research Blog Oct 12, 2026

Simulating student learning for better AI tutoring

StudentSim is a research system for evaluating and improving AI tutoring tools before they are tested with learners. It models how individuals may engage with tutors, giving feedback to improve tutor performance across subjects like chess, writing, and mathematics. The post Simulating student learning for better AI tutoring appeared first on Microsoft Research.

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.