FlowingFlowing
← Back to all development cases
Development case · blind model judgment pending independent human review; not a formal Benchmark conclusion.
AutocompleteWith cards better141 / 162 · c9b9dec2f15072ca

The brain as a blueprint: a survey of brain-inspired approaches to learning in artificial intelligence

Psychology and cognitive neuroscience · 2511.04455v1

FLOWING EVIDENCE BENCHMARK

How do we tell whether Evidence helps?

At the same writing position, with the same model and task, how does supplying retrieved paper passages change the first output? We compare matched versions and retain ties, unusable outputs and incomplete reviews.

SAME DRAFT · EVIDENCE ON OR OFF

01 · FIXED WRITING POSITION

MANUSCRIPT

Same writing position ▌

Matched autocomplete at the same draft position

02 · TWO MATCHED INPUTS

SHARED BY BOTH

Manuscript context, model, task and prompt

A · With Evidence

Retrieved paper passages supplied

B · Without Evidence

No retrieved passages supplied

LLM

Same model and version

A → first output

B → first output

Blind judge agent

First outputs are anonymized as X and Y

Continuation review: accuracy, fit to the writing task and usability
Returns: X preferred / tie / Y preferred / both unusable

Order check: X / Y → Y / X

HOW DOES BLIND JUDGING WORK?

① Anonymize both outputs
The judge sees the same draft and both first outputs without knowing which received Evidence.

② Compare and swap order
The judge applies task-specific criteria in X/Y and then Y/X order.

③ Review disagreements
A third pass resolves disagreements. Incomplete reviews remain in the denominator.

How are positions and cards selected?

The evaluation uses seven fields with forty fixed ordinary autocomplete positions in each, for 280 total. A position does not need to require a literature citation in its next sentence. Each Flowing request receives three to five retrieved cards: five when all pass the relevance gate, without padding with weak cards. The run supplied 1,355 cards: 244 positions received five, 27 received four, and nine received three.

ADOPTION · 150/280

Adoption coverage

After drafting, the model returns an evidenceUse array identifying cards that materially affected facts, correction, direction, scope, terminology, or expression. A nonempty array counts as adoption.

OUTCOMES · 70 / 7 / 30

Full outcomes after adoption

Paired outcomes are counted only when the model reports adoption and both arms produce text. Among 107 comparable positions, Flowing was preferred in 70, 7 tied, and no-Evidence was preferred in 30; one tie was unusable in both arms.

How are the two table percentages calculated?

ADOPTION COVERAGE

150adopted
÷
280all positions
=
54%

FLOWING PREFERENCE AFTER ADOPTION

70Flowing preferred
÷
107comparable adopted positions
=
65%

evidenceUse is the model’s structured report of its own generation process. It does not require a citation and is not rejected merely because a faithful source paraphrase is not verbatim. Actual manuscript citations continue to use separate strict source verification. The previous 140-position result remains retained for traceability.

These are development-stage model-reported adoption and blind-judge results pending independent human review. They are not formal Benchmark conclusions and do not, on their own, establish causality.

Manuscript writing position

Text excerpt · not a PDF page

Research manuscript · excerpt

The brain as a blueprint: a survey of brain-inspired approaches to learning in artificial intelligence

1 Introduction

…ndard deep learning models rely on biologically implausible error propagation algorithms and struggle to accumulate knowledge incrementally. While, the precise learning rule governing synaptic plasticity in biological systems remains unknown, recent discoveries in neuroscience could fuel further progress in AI. Here I examine successful implementations of brain-inspired principles in deep learning, current limitations, and promising avenues inspired by recent advances in neuroscience, including error computation, propagation, and integration via synaptic updates in biological neural networks.

↓ Next paragraph was generated here

The highlighted text is near the writing position. This is a layout of the evaluation input, not a PDF screenshot.

A matched comparison at the same writing positionJudgment: With cards better

Both arms use the same model, task and manuscript context. Only the retrieved passages differ. The first recorded output from each arm is shown below.

Evidence

Same model and position · retrieved passages supplied · first output

Preferred by blind review

A growing body of work highlights striking similarities between spiking neural networks that model biological networks and recurrent, binary neural networks, yet comparatively few studies examine the parallels between the learning dynamics of deep artificial neural networks and synaptic plasticity in spiking networks. This gap is largely attributable to the mismatch between the dynamical properties of synaptic plasticity and the requirements of gradient backpropagation.

Passages supplied to the Evidence version

Synaptic Plasticity Dynamics for Deep Continuous Local Learning (DECOLLE)

A growing body of work underlines striking similarities between spiking neural networks modeling biological networks and recurrent, binary neural networks. A relatively smaller body of work, however, discuss similarities between learning dynamics employed in deep artificial neural networks and synap…

Read full passage excerpt

A growing body of work underlines striking similarities between spiking neural networks modeling biological networks and recurrent, binary neural networks. A relatively smaller body of work, however, discuss similarities between learning dynamics employed in deep artificial neural networks and synaptic plasticity in spiking neural networks. The challenge preventing this is largely due to the discrepancy between dynamical properties of synaptic plasticity and the requirements for gradient backpropagation. Here, we demonstrate that deep learning algorithms that locally approximate the gradient backpropagation updates using locally synthesized gradients overcome this challenge. Locally synthesized gradients were initially proposed to decouple one or more layers from the rest of the network so as to improve parallelism.

Model-reported card use

The model did not report that any card materially affected this completion.

No Evidence

Same model and position · no retrieved passages · first output

Not preferred in this comparison

This survey is organized as follows. Section 2 reviews the biological foundations of synaptic plasticity and the error-propagation mechanisms that motivate brain-inspired learning rules. Section 3 then examines concrete implementations of these principles in deep learning architectures, before Section 4 discusses current limitations and Section 5 outlines promising avenues for future work.

Result of this comparison

The blinded model review preferred the Evidence version. Its original rationale appears below.

Read the original model judgment

Output B provides a coherent continuation grounded in the supplied source, directly incorporating and extending the DECOLLE paper's discussion of the gap between synaptic plasticity dynamics and gradient backpropagation requirements. It flows naturally from the draft's mention of 'error computation, propagation, and integration via synaptic updates in biological neural networks' by explaining why this gap exists. Output A, while structurally coherent, is a generic survey outline ('Section 2 reviews...Section 3 examines...') that does not engage with the specific source material or the draft's substantive focus on brain-inspired learning mechanisms. The draft already signals it will examine 'successful implementations...current limitations, and promising avenues,' making Output A's roadmap redundant rather than advancing the argument. Output B adds meaningful content about the specific challenge (dynamical properties mismatch) that the draft's neuroscience-inspired approach must address.

The passages are input material; their presence does not mean an output used them or that they support every claim. Draft and source passages are excerpted; no simulated scores or PDF appear here.