Anna-CPsyCounD

### 数据集概述
本数据集专为AnnaAgent智能体系统设计，模拟真实心理咨询场景下的虚拟来访者配置信息。数据集包含4个核心字段（portrait/id/report/conversation），涵盖用户画像、历史咨询报告及多轮对话记录，支持动态进化与多会话记忆模块的训练与评估。数据基于CPsyCounD数据集扩展生成，通过GPT-4o保证语义多样性与专业性。

### 字段说明
1. **portrait**（用户画像）
JSON结构，包含年龄、性别、婚姻状况、职业及症状等基础信息，用于初始化智能体状态。
示例：`{"age":"30","gender":"女","symptoms":"感情困惑;孤独感"}`

2. **id**（唯一标识符）
UUID格式，关联同一用户的多轮咨询数据，支持跨会话记忆检索。

3. **report**（历史报告）
结构化咨询记录，包含：
- 咨询经过（多轮对话摘要）
- 案例标题/类别（婚姻咨询/个人成长等）
- 经验感想（咨询师视角的总结）
- 运用的技术（心理咨询方法论）

4. **conversation**（对话记录）
JSON数组形式的多轮对话，包含角色（Seeker/Counselor）与内容，用于训练情感调制与主诉生成模块。

### 应用场景
- **动态进化模块**：通过report与conversation训练情绪调制器、主诉链生成器。
- **多会话记忆模块**：基于portrait与id实现跨会话状态分析与记忆检索。
- **心理咨询仿真**：模拟真实咨访交互，验证智能体在婚姻咨询等场景中的表现。

### 设计亮点说明
1. **记忆增强特性**：通过report与id字段支持长期记忆存储与跨会话检索，符合智能体"经验积累-知识提取"的核心能力。
2. **动态演化支持**：conversation字段包含情绪变化轨迹，可训练情绪调制器与主诉切换模块，呼应"环境动态交互"需求。
3. **隐私合规性**：采用GPT-4o生成合成数据，规避真实医疗数据泄露风险，同时保留专业咨询逻辑。

#### 下载方法
:modelscope-code[]{type="sdk"}
:modelscope-code[]{type="git"}

### Dataset Overview
This dataset is designed for the AnnaAgent AI Agent system, simulating virtual client configuration information in real-world psychological counseling scenarios. The dataset contains 4 core fields (portrait/id/report/conversation), covering user portraits, historical consultation reports and multi-round conversation records, supporting the training and evaluation of dynamic evolution and multi-session memory modules. The data is extended and generated based on the CPsyCounD dataset, and GPT-4o is used to ensure semantic diversity and professionalism.

### Field Descriptions
1. **portrait** (User Portrait)
JSON structure containing basic information such as age, gender, marital status, occupation and symptoms, used to initialize the agent's state.
Example: `{"age":"30","gender":"female","symptoms":"relationship confusion; loneliness"}`

2. **id** (Unique Identifier)
UUID format, used to associate multi-round consultation data of the same user and support cross-session memory retrieval.

3. **report** (Historical Consultation Report)
Structured consultation records, including:
- Consultation process (multi-round conversation summary)
- Case title/category (marital counseling/personal growth, etc.)
- Empirical reflections (summary from the counselor's perspective)
- Applied techniques (psychological counseling methodologies)

4. **conversation** (Conversation Record)
Multi-round dialogue in JSON array format, containing roles (Seeker/Counselor) and content, used to train emotion modulation and chief complaint generation modules.

### Application Scenarios
- **Dynamic Evolution Module**: Train emotion modulators and chief complaint chain generators through report and conversation data.
- **Multi-session Memory Module**: Implement cross-session state analysis and memory retrieval based on portrait and id fields.
- **Psychological Counseling Simulation**: Simulate real counselor-client interactions to verify the performance of AI Agents in scenarios such as marital counseling.

### Design Highlights
1. **Memory Enhancement Feature**: Supports long-term memory storage and cross-session retrieval through report and id fields, which aligns with the core capability of AI Agents' "experience accumulation - knowledge extraction".
2. **Dynamic Evolution Support**: The conversation field contains emotional change trajectories, which can be used to train emotion modulators and chief complaint switching modules, responding to the demand for "dynamic environmental interaction".
3. **Privacy Compliance**: Synthetic data generated by GPT-4o avoids the risk of real medical data leakage while retaining professional counseling logic.

#### Download Method
:modelscope-code[]{type="sdk"}
:modelscope-code[]{type="git"}

搜集汇总

背景与挑战

背景概述

Anna-CPsyCounD是一个用于模拟心理咨询场景的数据集，包含用户画像、历史咨询报告和多轮对话记录，支持动态进化与多会话记忆模块的训练与评估。数据基于CPsyCounD扩展生成，通过GPT-4o保证语义多样性与专业性，适用于AI心理健康领域的心理咨询仿真和多轮对话训练。