# MANAR: Memory-augmented Attention with Navigational Abstract Conceptual Representation

Source: [MANAR: Memory-augmented Attention with Navigational Abstract Conceptual Representation](<https://www.sota2.com/research/paper/69bcd49d0636e5bf4afc2a94>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-21T05:00:04.296Z

Authors: Zuher Jahshan, Ben Ben Ishay, Leonid Yavits

Published: 2026-03-19T09:37:18.000Z

## Abstract

MANAR (Memory-augmented Attention with Navigational Abstract Conceptual Representation), contextualization layer generalizes standard multi-head attention (MHA) by instantiating the principles of Global Workspace Theory (GWT). While MHA enables unconstrained all-to-all communication, it lacks the functional bottleneck and global integration mechanisms hypothesized in cognitive models of consciousness. MANAR addresses this by implementing a central workspace through a trainable memory of abstract concepts and an Abstract Conceptual Representation (ACR). The architecture follows a two-stage logic that maps directly to GWT mechanics: (i) an integration phase, where retrieved memory concepts converge to form a collective "mental image" (the ACR) based on input stimuli; and (ii) a broadcasting phase, where this global state navigates and informs the contextualization of individual local tokens. We demonstrate that efficient linear-time scaling is a fundamental architectural byproduct of instantiating GWT functional bottleneck, as routing global information through a constant-sized ACR resolves the quadratic complexity inherent in standard attention. MANAR is a compatible re-parameterization of MHA with identical semantic roles for its projections, enabling knowledge transfer from pretrained transformers via weight-copy and thus overcoming the adoption barriers of structurally incompatible linear-time alternatives. MANAR enables non-convex contextualization, synthesizing representations that provably lie outside the convex hull of input tokens - a mathematical reflection of the creative synthesis described in GWT. Empirical evaluations confirm that MANAR matches or exceeds strong baselines across language (GLUE score of 85.1), vision (83.9% ImageNet-1K), and speech (2.7% WER on LibriSpeech), positioning it as an efficient and expressive alternative to quadratic attention.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Automatic Speech Recognition on LibriSpeech (dev-clean)](<https://www.sota2.com/research/sota/automatic-speech-recognition-on-librispeech-dev-clean.md>): Automatic Speech Recognition; LibriSpeech (dev-clean); WER (%): 2.

- [Image Classification on ImageNet-1K (Top-1 Accuracy)](<https://www.sota2.com/research/sota/image-classification-on-imagenet-1k-top-1-accuracy.md>): Image Classification; ImageNet-1K; Top-1 Acc: 83.9.

- [Natural Language Understanding on GLUE (dev)](<https://www.sota2.com/research/sota/natural-language-understanding-on-glue-dev.md>): Natural Language Understanding; GLUE (dev); SST-2 (Acc): 94.4.

- [Automatic Speech Recognition on LibriSpeech (dev-other)](<https://www.sota2.com/research/sota/automatic-speech-recognition-on-librispeech-dev-other.md>): Automatic Speech Recognition; LibriSpeech (dev-other); WER: 6.4.

- [Automatic Speech Recognition on LibriSpeech clean (test)](<https://www.sota2.com/research/sota/automatic-speech-recognition-on-librispeech-clean-test.md>): Automatic Speech Recognition; LibriSpeech clean (test); WER: 2.7.

- [Automatic Speech Recognition on LibriSpeech (test-other)](<https://www.sota2.com/research/sota/automatic-speech-recognition-on-librispeech-test-other.md>): Automatic Speech Recognition; LibriSpeech (test-other); WER: 6.4.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2603.18676>)

[PDF](<https://arxiv.org/pdf/2603.18676.pdf>)
