FlowingFlowing
← Back to all development cases
Development case · blind model judgment pending independent human review; not a formal Benchmark conclusion.
AutocompleteWithout cards better146 / 162 · 5bc103a1a903e6ca

A Survey on Improving Human Robot Collaboration through Vision-and-Language Navigation

Psychology and cognitive neuroscience · 2512.00027v1

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

A Survey on Improving Human Robot Collaboration through Vision-and-Language Navigation

8. Conclusion and Future Directions > 8.4. Sim-to-Real Transfer Challenge

…uracy, as current models often fail to interpret vague commands when multiple candidates exist. Future work should enhance NLU by incorporating contextual reasoning and interactive clarification . By enabling robots to ask targeted questions and confirm user intent, such systems can resolve uncertainties and achieve more reliable, human-like navigation. In other words, key components include ambiguity resolution, contextual reasoning, and decentralized decision-making with dynamic role assignment. These capabilities support scalable and efficient robot coordination in real-world environments.

Coordinating task allocation among multiple robots grows increasingly complex as team size expands, yet most VLN research remains single-agent. This gap hinders consensus-building, conflict resolution, and resource optimization. Consider, for example, warehouse robots that collide or duplicate work due to poor communication. Future work should explore VLN-driven, decentralized frameworks for dynamic role assignment, enabling robots to share intentions, distribute tasks efficiently, and resolve conflicts collaboratively.

↓ 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: Without 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

Not preferred in this comparison

No continuation text was produced.

Model action：complete

Passages supplied to the Evidence version

A Coalition-Based Communication Framework for Task-Driven Flying Ad-Hoc Networks

The proposed coalition-based model provides a new perspective to the resource optimization of task-driven UAVs, making the networking more flexible. Due to the heterogeneity and task complexity, more complex factors and characteristics will exist in the optimization of UAV communication networks:

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

Preferred by blind review

Sim-to-real transfer remains a persistent obstacle for vision-and-language navigation, since policies trained in photorealistic simulators must contend with sensor noise, lighting variation, and dynamics that synthetic environments rarely reproduce faithfully. Closing this gap will likely require domain randomization, photorealistic rendering, and targeted fine-tuning on limited real-world trajectories.

Result of this comparison

The blinded model review preferred the no-Evidence version. Its original rationale appears below.

Read the original model judgment

Output A provides a coherent academic paragraph that continues the manuscript's discussion of future work challenges in VLN/robotics, maintaining thematic continuity with the draft's focus on real-world deployment obstacles (following paragraphs about NLU enhancement and multi-robot coordination). The sim-to-real transfer topic is a natural extension of the manuscript's trajectory from single-agent limitations toward practical deployment concerns. Output B is empty when the task clearly requires content completion, making it unusable.

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.