# Pro$^2$ Agent: Scientific Problem Probing for Autonomous AI R&D

> Pro²Agent's explicit Problem State—tracking diagnosed bottlenecks, candidate solutions, and evidence—achieves a record 76% medal rate on MLE-Bench, outperforming code-centric agents.

- **Source:** [arXiv](https://arxiv.org/abs/2611.25357)
- **Published:** 2026-11-21
- **Permalink:** https://picx.dev/p/p24yAW
- **Whiteboard:** https://picx.dev/p/p24yAW/image

## Summary

## Summary (Overview)

- **Core Contribution**: Pro²Agent introduces the **Problem State**, an explicit representation of what problems limit AI research performance, which candidate solutions likely address them, and the supporting evidence/confidence—shifting the primary state of autonomous AI research from a codebase state to a problem-centered state.
- **State-of-the-Art Results**: Achieves **76.0% medal rate** on the complete MLE-Bench with a 48-hour budget and **70.7%** with the standard 24-hour budget, the highest reported on the complete benchmark.
- **Transfers Across Domains**: Achieves competitive performance on PostTrainBench (30.69 weighted average), comparable to Claude Code powered by Opus 4.7, despite using the lightweight Gemini 3.5 Flash as the implementation model.
- **Stronger Initial Solutions**: On MLE-Bench Lite, Pro²Agent's first complete solution reaches a 57.6% Any Medal Rate (vs. 28.8% for MARS), with a normalized bronze gap of −0.57 (vs. 0.96 for MARS), meaning the average first solution starts inside the medal region.
- **Efficient Convergence**: Reaches medal-level performance within 5 high-quality scientific iterations, while MARS typically evaluates more than 70 repository variants without consistently reaching comparable scores.

## Introduction and Theoretical Foundation

### Background and Motivation

Large language model (LLM) agents can now autonomously conduct AI research by writing code and running experiments. However, existing agents typically formulate model development as **search or optimization over candidate code solutions**, iteratively optimizing a codebase while ignoring the problem structure behind the task. This formulation works on simpler tasks but struggles on difficult problems because:

- The agent never explicitly represents its understanding of *why* the task remains difficult
- The agent may repeatedly propose locally reasonable code changes while staying within the *wrong solution region*
- Experimental results are used only to accept/reject code changes, without updating a higher-level explanation of the problem

### Key Insight

> "A codebase state is necessary for autonomous AI research, but it provides only a partial view of the research process. It records what has been implemented, how the current solution performs, and which modifications have already been attempted. However, it does not explicitly capture the agent's evolving understanding of why the task remains difficult."

### Theoretical Foundation

The core theoretical contribution is the formalization of the **Problem State** as a nested structure:

$$S_t = \left\{ \left( P_i, E^p_i \right), \left\{ C_{i,j}, b_{i,j}, E^c_{i,j} \right\}_{j=1}^{L_i} \right\}_{i=1}^{K}$$

where:
- $P_i$ = the $i$-th identified performance-limiting problem
- $E^p_i$ = supporting evidence from diagnostic profiling
- $C_{i,j}$ = the $j$-th candidate solution for problem $P_i$
- $b_{i,j} \in [0,1]$ = dynamic belief confidence score
- $E^c_{i,j}$ = solution evidence set from literature or empirical traces

This contrasts with traditional code-space optimization:

$$m^* = \arg\max_{m \in \mathcal{M}} h(m)$$

where $\mathcal{M}$ is the space of valid code implementations and $h(\cdot)$ is a performance metric.

## Methodology

### Three-Phase Framework

#### Phase 1: Problem State Initialization
- **Dynamic runtime profiling**: Interactive execution probes inside a sandbox extract processing latencies, memory footprints, and domain-specific structural signals (including visual diagnosis for multimodal tasks)
- **Bottleneck-conditioned literature review**: Maps identified physical constraints to targeted web/academic search queries, retrieving state-of-the-art architectures and augmentations
- **Modular baseline construction**: Builds $M_0$ under strict security and profiling constraints, with validation evaluation decoupled as an HTTP grading service to prevent target leakage

#### Phase 2: Scientific Exploration Phase

Organized as a resource-bounded best-first search over a tree of hypothesis branches. Key components:

**Hypothesis Prioritization** (frequency-penalized score):
$$\text{Pri}(H_t) = \frac{b_{k,m}}{N(P_k) + 1}$$

where $b_{k,m}$ is the belief score and $N(P_k)$ counts historical exploration attempts targeting problem $P_k$. This penalizes repetitive failures and forces broader search.

**Belief Update Logic**:
- **Supported Hypothesis** ($\Delta h \geq \delta_{target}$): $b_{k,m} \leftarrow \min(1, b_{k,m} + \Delta b_{pos})$
- **Falsified/Unstable Hypothesis** ($\Delta h < \delta_{target}$ or crash): $b_{k,m} \leftarrow \max(0, b_{k,m} - \Delta b_{neg})$

**Single-Variable Control**: Strictly enforced at both code and architecture levels, with a Reviewer Agent rejecting any updates introducing confounding variables.

#### Phase 3: Engineering Optimization Phase

Models hyperparameter tuning as **Monte Carlo Tree Search (MCTS)** with:

**Audit Firewall** (pipeline invariance):
$$\text{C}_{\text{Diff}}(M_{root}, M_{child}) \cap \{\text{model}, \text{algo}, \text{data}\} = \emptyset$$

**Cost-Aware Reward Function**:
$$R(N_k) = \Delta h(N_k) - \gamma \ln\left(T_{exec}(N_k)\right)$$

where $\Delta h(N_k) = h(M_k) - h(M_{root})$ is the validation performance gain, $T_{exec}(N_k)$ is execution duration in seconds, and $\gamma > 0$ is a calibrated penalty weight.

### Tripartite Multi-Agent Loop

Three specialized roles ensure reliable implementation:

| Agent | Role | Model |
|-------|------|-------|
| **PI Agent** ($A_{PI}$) | Supervisor maintaining Problem State, formulating goals, inspecting trajectories every 10 steps | Gemini 3.1 Pro |
| **Implementation Agent** ($A_{imp}$) | Engineer in sandboxed environment, ReAct loop with file/diagnostic tools | Gemini 3.5 Flash |
| **Reviewer Agent** ($A_{rev}$) | Code auditor enforcing single-variable isolation (Phase 2) and guarding core modules (Phase 3) | Gemini 3.5 Flash |

Additional components include a dynamic budgeting protocol with an automated monitoring sentinel and an **Experience Bank** ($E_{bank}$) for persistent long-term memory with `retrieve_experience` and `write_experience` operations.

## Empirical Validation / Results

### MLE-Bench (Complete 75-Task Benchmark)

**Table 1: Performance comparison on complete MLE-Bench**

| Agent | Backbone LLM | Compute | Valid | Above Median | Bronze | Silver | Gold | Any Medal |
|-------|-------------|---------|-------|--------------|--------|--------|------|-----------|
| ML-Master | DeepSeek-R1 | A100, 12h | 93.3 | 44.9 | 4.4 | 7.6 | 17.3 | 29.3 |
| R&D-Agent | GPT-5 | V100, 12h | 53.3 | 40.4 | 6.7 | 12.0 | 16.4 | 35.1 |
| InternAgent | DeepSeek-R1 | A800, 12h | 96.4 | 48.4 | 7.1 | 10.7 | 18.7 | 36.4 |
| Famou-Agent | Gemini 2.5 Pro | A800, 24h | 96.9 | 51.6 | 8.4 | 12.4 | 22.7 | 43.6 |
| Kapso | Gemini 3 Pro Preview | –, 24h | – | – | – | – | – | 50.7 |
| ML-Master 2.0 | DeepSeek-V3.2-Speciale | 2×4090, 24h | 95.6 | 63.1 | 11.1 | 25.8 | 19.6 | 56.4 |
| MARS | Gemini 3 Pro Preview | A100, 24h | 98.7 | 65.8 | 9.3 | 15.6 | 31.1 | 56.0 |
| MARS+ | Gemini 3 Pro Preview | 2×H100, 24h | 100.0 | 74.2 | 12.4 | 16.4 | 33.8 | 62.7 |
| AIBuildAI | Claude Opus 4.6 | A100, 24h | 100.0 | 71.1 | – | – | 25.8 | 63.1 |
| Famou-Agent 2.0 | Gemini 3 Pro Preview | A800, 24h | – | – | – | – | – | 64.4 |
| Iris | Claude Opus 4.6 | –, 12h | – | – | – | – | – | 64.9 |
| MLEvolve | Gemini 3.1 Pro Preview | H200, 12h | 100.0 | 76.0 | – | – | 34.7 | 65.3 |
| ScienceFlow | DeepSeek-V4-Flash-Preview | ≤2 GPUs, 24h | – | – | – | – | – | 70.2 |
| AIBuildAI-2 | Claude Opus 4.7 | A100, 24h | – | – | – | – | – | 70.7 |
| **Pro²Agent** | **Gemini 3.5 Flash‡** | **A100, 24h** | **100.0** | **84.0** | **25.3** | **21.3** | **24.0** | **70.7** |
| **Pro²Agent** | **Gemini 3.5 Flash‡** | **A100, 48h** | **100.0** | **84.0** | **29.3** | **21.3** | **25.3** | **76.0** |

### PostTrainBench (28 Configurations)

**Table 2: Performance on complete PostTrainBench**

| Method | Avg. | AIME 2025 | ArenaHard | BFCL | GPQA | GSM8K | HealthBench | HumanEval |
|--------|------|-----------|-----------|------|------|-------|-------------|-----------|
| Official Instruct Models | 51.14 | 29.17 | 70.21 | 85.00 | 36.21 | 87.00 | 43.32 | 71.49 |
| Fable 5 (Claude Code) | 41.79±1.7 | 13.33±2.4 | 61.46±10.8 | 73.00±32.9 | 28.35±2.7 | 83.53±1.9 | 40.12±9.4 | 58.38±13.8 |
| GPT 5.6 (Sol, Codex) | 36.23±0.2 | 7.50±3.5 | 26.40±13.5 | 94.38±1.2 | 30.13±1.2 | 69.48±2.7 | 27.39±3.4 | 63.95±2.7 |
| Opus 4.7 (Claude Code) | 28.56±1.7 | 6.39±3.2 | 19.59±18.1 | 76.91±28.0 | 27.48±4.0 | 60.72±17.0 | 17.81±8.7 | 42.07±22.9 |
| Gemini 3.1 Pro (OpenCode) | 21.99±1.8 | 3.89±1.9 | 7.42±5.4 | 63.59±27.7 | 18.53±8.3 | 45.51±22.3 | 16.33±9.8 | 40.19±8.4 |
| **Pro²Agent (Gemini 3.5 Flash)‡** | **30.69±3.1** | **5.75±4.4** | **28.59±12.2** | **78.60±17.8** | **26.45±3.3** | **64.47±5.1** | **19.22±6.9** | **50.63±16.6** |

### Controlled Studies on MLE-Bench Lite

**Table 3: Initial solution quality and node efficiency**

| Method | Above Median | Any Medal | Gold | $G_{bronze}$ ↓ | Build Time (min) | Medal@3 | Medal@5 | Node AUC ↑ |
|--------|-------------|-----------|------|----------------|------------------|---------|---------|------------|
| Direct Coding Loop | 25.8±1.5 | 12.1±3.0 | 4.5±0.0 | 1.56±0.09 | 2.1±1.2 | 21.2±3.0 | 30.3±3.0 | −0.60±0.07 |
| Search-Augmented | 39.4±3.0 | 28.8±4.0 | 12.1±1.5 | 1.23±0.07 | 8.4±2.3 | 30.3±1.5 | 36.4±5.5 | −0.44±0.12 |
| MARS | 50.0±2.6 | 28.8±4.0 | 12.1±1.5 | 0.96±0.20 | 14.8±2.6 | 34.8±3.0 | 39.4±5.5 | −0.39±0.09 |
| **Pro²Agent** | **81.8±2.6** | **57.6±1.5** | **27.3±0.0** | **−0.57±0.05** | **11.2±4.4** | **63.6±2.6** | **66.7±1.5** | **0.58±0.17** |

The normalized bronze gap is defined as:

$$G_{bronze}(s_t) = \frac{b_t - s_t}{|b_t - m_t| + \epsilon}$$

where $b_t$ and $m_t$ denote the bronze threshold and leaderboard median, and $G_{bronze} < 0$ indicates medal-level performance.

### Ablation Study

**Table 4: Ablation of the Problem State**

| Variant | Initial Medal | Final Medal | Node AUC ↑ |
|---------|--------------|-------------|------------|
| Codebase State | 10.6±1.5 | 56.1±3.0 | 0.06 |
| Prob. State w/o Grounding | 47.0±1.5 | 77.3±2.6 | 0.27 |
| Static Prob. State | 56.1±1.5 | 83.3±3.0 | 0.45 |
| **Full Pro²Agent** | **57.6±1.5** | **89.4±1.5** | **0.66** |

### Three-State Bottleneck Taxonomy

Across diverse trajectories, Pro²Agent's Problem State provides a structured taxonomy of research outcomes:

1. **Resolved Bottleneck**: A diagnosed bottleneck is experimentally mitigated enough that it is no longer the dominant limitation (e.g., `bms-molecular-translation`, `gemma3-4b-gsm8k`)
2. **Resource-Limited Bottleneck**: The bottleneck is correctly diagnosed and supported by literature, but the intervention cannot be trained to convergence within available compute (e.g., `rsna-breast-cancer-detection`)
3. **Unresolved Bottleneck**: Experiments narrow the likely failure source but do not establish an effective intervention (e.g., `qwen3-1.7b-aime`)

## Theoretical and Practical Implications

### Theoretical Implications

- **Problem State as First-Class State**: The paper demonstrates that representing understanding of *why* a task is difficult, rather than just *what* code has been tried, fundamentally changes how autonomous research agents explore solution spaces.
- **Experimentation as Hypothesis Testing**: Instead of using experiments only to accept/reject code changes, Pro²Agent uses experiments to test and revise its understanding of the problem—turning local code improvements into a directed research process.
- **Evidence-Grounded Belief Revision**: The explicit separation of problems, candidate solutions, and evidence enables principled belief updates that adapt to new experimental outcomes.

### Practical Implications

- **Stronger Initial Solutions**: Grounding design choices in explicit problem understanding rather than LLM preferences produces substantially better first solutions (57.6% vs. 28.8% medal rate on first solution).
- **Compute Efficiency**: Hypothesis-guided iteration concentrates compute on scientifically distinct decisions rather than many local code modifications, reaching medal-level performance within 5 iterations vs. 70+ for MARS.
- **Model Efficiency**: The tripartite architecture uses a lightweight model (Gemini 3.5 Flash) for implementation while reserving a stronger model (Gemini 3.1 Pro) for high-level diagnosis—achieving results comparable to systems using frontier models like Claude Opus 4.7.
- **Transparency and Reusability**: Each run produces an inspectable record of diagnosed bottlenecks, tested hypotheses, and outcomes, enabling human researchers to audit and reuse the research process.

## Conclusion

Pro²Agent introduces a paradigm shift in autonomous AI research: treating the agent's understanding of the problem as a first-class research state. Instead of representing progress only through the current codebase, evaluation scores, and modification history, Pro²Agent maintains a Problem State that records performance-limiting problems, candidate solutions, supporting evidence, and confidence. This changes experimentation from direct code selection into a process for testing explanations of why performance is limited.

**Key findings**:
- On MLE-Bench: 76.0% medal rate (48h) and 70.7% (24h), the highest reported on the complete benchmark
- On PostTrainBench: Comparable to Claude Code with Opus 4.7 while using a lighter implementation model
- Ablations confirm that evidence grounding and continual Problem State updates contribute beyond static planning

**Future directions**:
1. Scaling Problem Probing to longer horizons and parallel hypothesis testing
2. Exploring other model allocations in the tripartite architecture
3. Richer belief revision mechanisms (probabilistic, noise-aware updates)
4. Reusable research knowledge across tasks, enabling agents that accumulate scientific understanding rather than only code

> "These findings suggest that progress in autonomous AI research depends not only on stronger language models or larger code-search budgets, but also on maintaining an explicit, evidence-grounded, and revisable understanding of the problem."

---

_Markdown view of https://picx.dev/p/p24yAW, served by PicX — AI-generated visual whiteboard summaries of research papers._
