讲座嘉宾简介
埃尔古诺娃·奥尔加·季托夫娜（Ergunova Olga Titovna），经济学博士，圣彼得堡彼得大帝理工大学副教授、生产管理高等学院成员。她是俄罗斯科学基金会第25-28-01469号项目“超大都市数字经济中管理社会劳动关系的神经网络解决方案”的负责人。埃尔古诺娃博士担任金砖国家女科学家协会执行主任、俄罗斯科学院专家、俄罗斯青年科学家联盟副主席。她在人工智能、数字转型、智慧城市、旅游与酒店业等领域拥有22年以上的管理与教学经验，出版论文300余篇，并担任多部国际学术著作的编辑。她也是多项国际青年科学家竞赛的组织者，并多次获得国家级奖项。
讲座（一）
讲座时间：4月18日13:00-17:00
讲座地点：管楼506会议室
讲座主题： Social Protection of Workers in the Platform Economy Digital Solutions and the Role of the State
讲座摘要
本讲座聚焦平台经济中劳动者的社会保护问题。全球平台经济市场规模预计2026年达6740亿美元，劳动者数量达1.54–4.35亿，但多数缺乏健康、养老、工伤等基本保障。讲座系统比较了主要经济体的监管模式，重点分析算法管理对劳动者权利的影响，以及女性、青年、残障人士和农村劳动者的脆弱性。讲座提出“便携式福利”系统——保护随劳动者而非雇主转移，并展示了如何利用LSTM预测、数字孪生和AI审计等数字技术支撑平台劳动政策的制定与评估。中俄可在按单保险、平台登记、纠纷解决等方面相互借鉴。
This lecture addresses the social protection deficit for workers in the platform economy. With the global gig economy reaching $674 billion in 2026 and 154–435 million workers, most lack access to health, pension, and injury insurance. The lecture compares regulatory models: the EU’s Directive 2024/2831 (presumption of employment), Spain’s Rider Law, China’s hukou reform and order-based injury insurance pilots, and Russia’s Federal Law No. 289-FZ (platform registry). It examines algorithmic management, gender and youth vulnerabilities, and the crisis of the classical employer-based protection model. A portable benefits system — where rights follow the worker across platforms — is proposed. Digital solutions such as LSTM forecasting, digital twins, and AI fairness auditing are introduced as tools for evidence-based policy. China and Russia can learn from each other in areas like per-order insurance, platform transparency, and dispute resolution.
讲座（二）
讲座时间：4月19日13:00-17:00
讲座地点：管楼506会议室
讲座主题：Cloud-Fog Architecture with Adaptive PID Feedback Control for LSTM-Based Labor Market Forecasting in Megacities
讲座摘要
本讲座介绍一种面向超大都市劳动力市场预测的智能架构，结合云计算、雾计算、LSTM深度学习和自适应PID反馈控制。传统预测模型难以应对超大都市就业数据的非线性、高波动和实时性需求。该研究提出三层架构：云端负责全局LSTM模型训练与超参数优化；雾端执行实时推理、PID误差校正与低延迟决策；边缘层采集招聘门户、社交媒体和政府统计数据。PID控制器根据预测残差动态调整比例、积分、微分增益，形成闭环反馈，显著降低预测漂移。预期成果包括将预测误差（RMSE）降低15–25%，雾端推理延迟低于500毫秒，并实现比开环模型快3倍的适应速度。该方法为城市劳动力治理提供了数据驱动的实时决策支持工具。
This lecture presents a hybrid cloud-fog computing architecture integrated with LSTM deep learning and adaptive PID feedback control for labor market forecasting in megacities. Traditional models fail to capture the non-linear, volatile, and real-time nature of urban employment data. The proposed three-tier system includes: a cloud tier for global LSTM training and hyperparameter optimization; a fog tier for real-time inference, PID error correction, and low-latency decision support; and an edge tier for data ingestion from job portals, social media, and government statistics. The adaptive PID controller dynamically adjusts proportional, integral, and derivative gains based on residual errors, forming a closed feedback loop to correct prediction drift. Expected outcomes include a 15–25% reduction in RMSE, sub-500ms fog-tier latency, and threefold faster adaptation than open-loop models. This framework enables evidence-based, real-time labor policy for megacities.
