# AI use cases for Effective Auto Research EAR (liyingze07-svg/EffectiveAutoResearch-EAR-)

Repo: [liyingze07-svg/EffectiveAutoResearch-EAR-](https://github.com/liyingze07-svg/EffectiveAutoResearch-EAR-) | Use cases: 4 | Page: https://dedupedai.com/use-cases/liyingze07-svg-effectiveautoresearch-ear/

EAR is an open Python research workbench that helps with idea discovery, autonomous mathematical research, and evidence-linked reviewer responses.

This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL.

- Explore the EAR demo in a local browser studio: https://dedupedai.com/use-cases/liyingze07-svg-effectiveautoresearch-ear/explore-ear-demo-in-browser/
- Turn a research direction into a ranked proposal: https://dedupedai.com/use-cases/liyingze07-svg-effectiveautoresearch-ear/turn-research-direction-into-ranked-proposal/
- Inspect saved proof artifacts from the offline math run: https://dedupedai.com/use-cases/liyingze07-svg-effectiveautoresearch-ear/inspect-offline-math-proof-artifacts/
- Draft evidence-linked replies to reviewer comments: https://dedupedai.com/use-cases/liyingze07-svg-effectiveautoresearch-ear/draft-evidence-linked-rebuttal/

## How to explore the EAR demo in a local browser studio with Effective Auto Research EAR

Repo: [liyingze07-svg/EffectiveAutoResearch-EAR-](https://github.com/liyingze07-svg/EffectiveAutoResearch-EAR-) | Difficulty: Intermediate | Time: 15 minutes | Basis: From the README

Run the offline studio to see idea trees, proof checks and rebuttal traces with no API key.

**Who it is for:** A researcher or evaluator who wants to see what EAR produces before setting up any model account.

**The problem:** Setting up live research tools takes time, and it is hard to judge the workbench from a repository page alone.

**The outcome:** A local studio page where you can search the idea tree, inspect a critique, download a proposal, check a proof and download a paper, and trace reviewer evidence.

### Steps

1. Install Python 3.10 or newer.
2. Run `git clone https://github.com/liyingze07-svg/EffectiveAutoResearch-EAR-.git EAR`
3. Run `cd EAR`
4. Run `python3 -m ear demo studio`
5. Open the studio page the command starts, select a module, and inspect or download its output.
6. Run `python3 -m ear demo check` to validate the idea trace and synthetic rebuttal wiring.

### Prompt to paste into your agent

```text
Clone the repo liyingze07-svg/EffectiveAutoResearch-EAR- into <target-folder>, then run `python3 -m ear demo check` from inside it. Report the output of the check, and state whether it validated the idea trace and synthetic rebuttal wiring. Do not claim any live model result, since the demo uses authored examples.
```

### Caveat

The demo uses authored examples, so its outputs show the workflow shape and are not evidence of live research results.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/liyingze07-svg-effectiveautoresearch-ear/explore-ear-demo-in-browser/

## How to turn a research direction into a ranked proposal with Effective Auto Research EAR

Repo: [liyingze07-svg/EffectiveAutoResearch-EAR-](https://github.com/liyingze07-svg/EffectiveAutoResearch-EAR-) | Difficulty: Advanced | Time: 120 minutes | Basis: From the README

Use AutoVibeIdea to map literature, rank candidate ideas and draft a proposal.

**Who it is for:** A researcher starting a new project who has a broad direction and a set of constraints but no settled idea yet.

**The problem:** Early candidate ideas often overlap with existing work, and checking that by hand takes a long time.

**The outcome:** A literature landscape, ranked candidates, and a proposal draft saved in its own AutoVibeIdea workspace, with critique that can rule out overlapping ideas early.

### Steps

1. Complete the live setup in docs/getting-started.md, including an authenticated Codex or Claude route and the AutoVibeIdea dependencies.
2. Write your research direction and constraints as the module input, following docs/getting-started.md#develop-a-research-direction.
3. Run the AutoVibeIdea workflow from the autovibeidea/ folder in its own workspace, using the exact command given in the linked guide. (inferred, check the README)
4. Review the literature landscape and the ranked candidates.
5. Inspect the critique for any candidate that shows overlap with prior work, then take the strongest proposal forward into a validation plan.

### Prompt to paste into your agent

```text
In the repo liyingze07-svg/EffectiveAutoResearch-EAR-, follow docs/getting-started.md to run AutoVibeIdea on the research direction <research-direction>. Expected output is a literature landscape, ranked candidates and a proposal draft in its own workspace. Verify that each candidate's novelty judgment cites retrieved prior work before you rank it highest.
```

### Caveat

The README says historical model judgments do not establish conference acceptance, so treat the ranking as a starting point for your own review.

Risk, credentials: Live runs use your own authenticated Codex or Claude account, so keep those credentials out of shared workspaces and commits.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/liyingze07-svg-effectiveautoresearch-ear/turn-research-direction-into-ranked-proposal/

## How to inspect saved proof artifacts from the offline math run with Effective Auto Research EAR

Repo: [liyingze07-svg/EffectiveAutoResearch-EAR-](https://github.com/liyingze07-svg/EffectiveAutoResearch-EAR-) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README

Run the offline math fixture to see proof checks, rejection and saved manuscript files.

**Who it is for:** A mathematician or developer who wants to see how the AutonomousMath engine records its work before using a live backend.

**The problem:** Live math runs need a configured backend, so it is hard to learn the artifact layout and review flow first.

**The outcome:** A workspace with a labeled offline demo run showing rejection, revision and saved math artifacts, with accepted set to false.

### Steps

1. Clone the repo and run `cd EAR` from the project root.
2. Run `python3 -m ear math run --offline --workspace workspaces/math-demo --episodes 1`
3. Run `python3 -m ear math status --workspace workspaces/math-demo`
4. Open the saved artifacts in workspaces/math-demo and confirm the run is labeled offline_demo with accepted=false.
5. Read autonomousmath/docs/ENGINE_DESIGN.md to map each saved artifact to its role in the review gate.

### Prompt to paste into your agent

```text
In the repo liyingze07-svg/EffectiveAutoResearch-EAR-, run `python3 -m ear math run --offline --workspace workspaces/math-demo --episodes 1` and then `python3 -m ear math status --workspace workspaces/math-demo`. Report the status output and the saved artifact list. Verify that the run is labeled offline_demo and accepted=false before saying anything about its contents.
```

### Caveat

The offline fixture is a demonstration only, so its output is not a completed research result and does not show what a live backend would produce.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/liyingze07-svg-effectiveautoresearch-ear/inspect-offline-math-proof-artifacts/

## How to draft evidence-linked replies to reviewer comments with Effective Auto Research EAR

Repo: [liyingze07-svg/EffectiveAutoResearch-EAR-](https://github.com/liyingze07-svg/EffectiveAutoResearch-EAR-) | Difficulty: Advanced | Time: 120 minutes | Basis: From the README

Use AutoRebuttal to tie each reviewer concern to paper evidence before writing replies.

**Who it is for:** An author preparing a rebuttal who has a paper and a set of reviewer comments and needs replies that point to evidence.

**The problem:** Replies written without checking the paper can overstate results or miss concerns, and the AC comment is easy to mix with the replies.

**The outcome:** Evidence-linked replies, an AC comment, and an unresolved-concern ledger kept as separate outputs.

### Steps

1. Gather the paper and the reviewer comments as module inputs, as described in rebuttal/.
2. Complete the live setup in docs/getting-started.md, following docs/getting-started.md#prepare-a-rebuttal-case.
3. Run the AutoRebuttal workflow from the rebuttal/ folder, using the exact command given in the linked guide. (inferred, check the README)
4. Check each reply against the paper evidence it cites, and check the staged review results in the rebuttal/ outputs.
5. Keep the replies, the AC comment and the unresolved-concern ledger in separate files before you submit anything.

### Prompt to paste into your agent

```text
In the repo liyingze07-svg/EffectiveAutoResearch-EAR-, follow docs/getting-started.md to prepare a rebuttal for the paper <paper-path> and reviewer comments <comments-path>. Expected output is evidence-linked replies, an AC comment and an unresolved-concern ledger as separate files. Verify that every reply cites a specific passage or result in the paper before you report it as supported.
```

### Caveat

The outputs are drafts that still need author review, since the README describes staged checks but does not guarantee that a reply will persuade a reviewer.

Risk, credentials: Live rebuttal runs use your own authenticated model account, so do not paste account credentials into the paper or comment inputs.

Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/

URL: https://dedupedai.com/use-cases/liyingze07-svg-effectiveautoresearch-ear/draft-evidence-linked-rebuttal/