活动期间将组织专场对接会，为技术供需双方提供独立的洽谈空间和良好的交流研讨环境，搭建校企技术咨询和技术对接平台。作为系列活动之一，“2021沪苏同城国际创新挑战赛专场对接会“也将在活动期间同步举行。

During the event, a special matchmaking meeting will be organized to provide independent negotiation space and a good exchange and discussion environment for both technology suppliers and demanders, and build a school-enterprise technical consultation and technology docking platform. As one of the series of activities, the "2021 Shanghai and Suzhou City International Innovation Challenge Special Matchmaking Meeting" will also be held simultaneously during the event.

需求方
Demand side		 需求名称
Demand name		 需求内容
Demand content
中国科学院电子学研究所苏州研究院
Suzhou Research Institute, Institute of Electronics, Chinese Academy of Sciences		 突发事件实时检测、事件知识表示与潜在因素挖掘、重大事件演化分析与突变预警等关键技术
Key technologies such as real-time detection of emergencies, event knowledge representation and potential factor mining, major event evolution analysis, and mutation warning		 针对社交媒体大数据中事件（如恐怖袭击、自然灾害等）信息分布稀 疏性、偏斜性问题，研究突发事件实时检测、事件知识表示与潜在因素挖 掘、重大事件演化分析与突变预警等关键技术，具备事件信息提取、事件 要素关联、突发事件检测、临近预警等能力。重大突发事检测时效性达到分钟级；事件检测平均准确率超过 90%，召回率超过 90%；事件预警平均准 确率超过 80%。
In response to the sparse and skewed information distribution of events (such as terrorist attacks, natural disasters, etc.) in social media big data. Research on key technologies such as real-time detection of emergencies, event knowledge representation and potential factor mining, major event evolution analysis and mutation warning. It has the capabilities of event information extraction, event element correlation, emergency detection, approaching early warning, etc. The timeliness of detection of major emergencies reaches the minute level. The average accuracy rate of event detection exceeds 90%, and the recall rate exceeds 90%. The average accuracy of event early warning exceeds 80%.
中国科学院电子学研究所苏州研究院
Suzhou Research Institute, Institute of Electronics, Chinese Academy of Sciences		 自动化数据标注技术
Automated data annotation technology		 针对自然语言处理领域数据标注人工成本高、耗时长等问题，研究自 动化数据标注技术，通过知识库或少量样本库构建远程监督模型，实现大 批量样本的自动化回标。数据标注类型不少于 3 种，包括实体、关系、属 性；数据标注人工参与度小于 10%，数据标注准确率大于 99%；基于自动化 标注生成的样本数据进行模型训练，与人工标注样本对比，模型性能没有下降。
Aiming at the problems of high labor cost and time-consuming data annotation in the field of natural language processing, research on automatic data annotation technology. Build a remote supervision model through a knowledge base or a small sample library to realize the automatic return of a large number of samples. There are no less than 3 types of data annotations, including entities, relationships, and attributes. The manual participation rate of data labeling is less than 10%, and the data labeling accuracy rate is greater than 99%. Model training is performed based on the sample data generated by automatic labeling. Compared with manual labeling samples, the model performance does not decrease.
苏州普元数字科技有限公司
Primeton Software, Inc. Suzhou		 政务知识图库
Government Affairs Knowledge Gallery		 1.对政务主题库中政务业务数据进行学习识别，通过学习算法得出其 业务知识内容，包括对选定业务数据样本空间的模式学习，不断的算法优 化，实现对业务数据的知识提取，并通过校核机制，对提取成果做校正， 形成知识识别提取算法、模型和机制；
2.对业务知识进行标签化归类、并自动识别其关联性，成知识点到知 识点的关联识别和自动创建关联关系；
3.基于拓扑图模式和算法，形成图结构的知识脉络数据，并落地至图 数据库中进行存储，提供可查询获取的数据服务；
4.以知识图谱形式对政务知识进行展示，支持图谱式展示和关联延展、 上下钻取等查看方式，支持知识检索。包括知识图谱展示门户和相关功能， 以可视化方式展示图谱，支撑知识图谱的拖拽延展展示、上下层级钻取等 多种展示方式、支持基于不同维度和粒度的知识检索和获取；
5.最终形成政务领域的业务知识积累库，为政务需求和社会需求提供 数据支撑，推动政务业务和应用创新。
1. Learning and identifying the government affairs business data in the government affairs theme database, and obtaining its business knowledge content through the learning algorithm. Including model learning of selected business data sample space, continuous algorithm optimization, and realization of knowledge extraction of business data. Through the verification mechanism, the extraction results are corrected to form knowledge recognition and extraction algorithms, models and mechanisms
2. Labeling and categorizing business knowledge and automatically identifying its relevance. Identify and automatically create associations from knowledge points to knowledge points
3. Based on the topological graph model and algorithm, form the knowledge network data of the graph structure. And land in the graph database for storage, and provide data services that can be queried and obtained
4. Display the government affairs knowledge in the form of knowledge graph, support graph display and related extension, drill up and down and other viewing methods, and support knowledge retrieval. Including the knowledge graph display portal and related functions, displaying the graph in a visual manner, supporting the drag and extension display of the knowledge graph, and the upper and lower levels of drilling and other display methods. Support knowledge retrieval and acquisition based on different dimensions and granularity
5. Eventually form a business knowledge accumulation library in the field of government affairs, provide data support for government affairs and social needs, and promote government affairs and application innovation
哈金森工业橡胶制品(苏州）有限公司
Hutchinson Industrial Rubber Products (Suzhou) Co., Ltd.		 Ai 人工智能检测
Ai artificial intelligence detection		 技术问题：传统的工业质量检验方法是通过放大镜与显微镜来手动检 测各项参数，但无法实现自适应，泛用性低，并且人工成本高。而 Ai 人工 智能检测可以实现对生产质量数据的完全自动控制，并为过程优化和过程 提供关键数据支持。
预期达到效果：可有效提升检测速度，延迟时间短，降低设备的消耗， 有效解决人力不足的缺陷。
Technical problem: The traditional industrial quality inspection method is to manually inspect various parameters through a magnifying glass and a microscope. However, self-adaptation cannot be achieved, versatility is low, and labor costs are high. Ai artificial intelligence detection can realize complete automatic control of production quality data and provide key data support for process optimization and process.
Expected results: It can effectively improve the detection speed, the delay time is short, the consumption of equipment is reduced, and the defect of insufficient manpower can be effectively solved.
迈杰转化医学研究（苏州）有限公司
MEDx (Suzhou) Translational Medical Co., Ltd		 人工智能（AI）病理诊断和远程病理诊断服务产品开发与应用
Development and application of artificial intelligence (AI) pathological diagnosis and remote pathological diagnosis service products		 人工智能（AI）病理诊断可以辅助病理医生阅片判读，提高病理医生 的工作效率，并且能够应付各种 Biomarker IHC 检测判读的特别需求（如 密度、具体细胞个数等）。
Artificial intelligence (AI) pathological diagnosis can assist pathologists in reading and interpretation, improve the work efficiency of pathologists, and can meet the special needs of various Biomarker IHC detection and interpretation (such as density, specific cell number, etc.).
中国科学院电子学研究所苏州研究院
Suzhou Research Institute, Institute of Electronics, Chinese Academy of Sciences		 基于人机混合的认知增强技术
Cognitive enhancement technology based on man-machine hybrid		 针对风险评估、趋势预测等重大领域中人工智能可解释性差、置信度 低、推理能力不足等问题，研究基于人机混合的认知增强技术，能够通过 专家-专家、专家-机器、机器-机器的多重协同交互，提升智能系统对开放、 模糊问题的认知和推断能力，并建立反映大规模动态逻辑关联体系的人机 协同分析认知视图。通过人机混合，机器推断过程转化为可解释逻辑关系 或规则的比率不低于 80%，对低置信预测结果（低于 0.7）的置信度平均提 升不小于 0.2。
Aiming at problems such as poor interpretability, low confidence, and insufficient reasoning ability of artificial intelligence in major fields such as risk assessment and trend prediction. The research is based on the human-machine hybrid cognitive enhancement technology, which can achieve multiple collaborative interactions of expert-expert, expert-machine and machine-machine. Improve the cognition and inference ability of intelligent systems on open and ambiguous problems. And establish a human-machine collaborative analysis cognitive view that reflects a large-scale dynamic logical association system. Through man-machine mixing, the ratio of machine inference processes into interpretable logical relationships or rules is not less than 80%, and the confidence of low-confidence prediction results (less than 0.7) is increased by not less than 0.2 on average.
中国科学院电子学研究所苏州研究院
Suzhou Research Institute, Institute of Electronics, Chinese Academy of Sciences		 面向互联网的行为体分析技术
Internet-oriented actor analysis technology		 针对实体空间中的人物、组织在互联网空间中的行为体，研究其行为 规律挖掘、行为行动预测与虚实关系映射技术，建立网络行为体关系图谱， 实现网络行为体社区发现、时空规律生成、立场预测、应激反应预测等。 行为体分析数据来源不少于 3 种，包括社交媒体、新闻、百科等；网络行 为体关系图谱节点数大于 100 万；针对新议题的立场预测准确率大于 80%。
Aiming at the characters in the physical space and the actors of the organization in the Internet space, research their behavior rule mining, behavior prediction and virtual-real relationship mapping technology. Establish the relationship graph of network actors, realize community discovery of network actors, generation of temporal and spatial laws, position prediction, stress response prediction, etc. There are no less than three sources of behavioral analysis data, including social media, news, and encyclopedias. The number of nodes in the network behavior relationship graph is more than 1 million; the accuracy of position prediction on new issues is more than 80%.
和铂医药（苏州）有限公司
Harbour BioMed (Suzhou) Co., Ltd.		 AI 技术助力抗体药物研发
AI technology helps antibody drug development		 利用 AI 技术，开发助力早期抗体发现的新方法，提高早期抗体发现的 效率和成功率。可以包括以下几个方面：1）AI 辅助构建具有更优生物学功 能和理化性质的抗体序列库，并开发新的算法，充分利用机器学习和分子 动力学，以实现对针对难以攻克靶点的特异性抗体的开发。2）运用 AI 模 型，通过对候选抗体分子序列的分析，预测抗体分子的功能及物理化学性 质，以减少完全依靠实验方法来进行筛选的工作量。3）AI 辅助抗体分子的 优化改造。如针对某些具有优良的功能，但是在物理化学特性上存在缺陷的抗体分子，利用 AI 的预测功能，指导抗体分子的改造和优化。最终目标 是加快抗体药物开发的速度和提高抗体的生物和理化特性。
Use AI technology to develop new methods to facilitate early antibody discovery and improve the efficiency and success rate of early antibody discovery. It can include the following aspects: 1) AI assists in the construction of antibody sequence libraries with better biological functions and physical and chemical properties. And develop new algorithms, make full use of machine learning and molecular dynamics to achieve the development of specific antibodies against difficult targets. 2) Use the AI model to predict the function and physicochemical properties of the antibody molecule by analyzing the sequence of the candidate antibody molecule, so as to reduce the workload of screening completely relying on experimental methods. 3) AI assists the optimization of antibody molecules. For example, for some antibody molecules that have excellent functions but have defects in physical and chemical properties. Use the predictive function of AI to guide the transformation and optimization of antibody molecules. The ultimate goal is to speed up the development of antibody drugs and improve the biological and physical and chemical properties of antibodies.
江苏力鼎环保装备有限公司
Jiangsu LiDing Environmental Protection Equipment Co., Ltd.		 先进污水处理技术
Advanced sewage treatment technology		 需要解决的技术问题： 1.国内外先进生活污水处理技术（总氮高效去除，非要求投加碳源）； 2.国内外先进处理技术（高效除磷，非加药、电解等）； 3.数字化软件开发（先进的 BIM\GIS 技术应用开发，应用农村管网收 集系统设计应用）。 技术指标（污染物平均去除率）： 总氮≥65%、总磷≥88%。
Technical problems to be solved: 1. Domestic and foreign advanced domestic sewage treatment technology (high-efficiency removal of total nitrogen, non-required addition of carbon source); 2. Domestic and foreign advanced treatment technology (high-efficiency phosphorus removal, non-dosing, electrolysis, etc.); 3 .Digital software development (advanced BIM\GIS technology application development, application of rural pipe network collection system design and application). Technical indicators (average removal rate of pollutants): total nitrogen ≥65%, total phosphorus ≥88%.
苏州久之新能源科技有限公司
Fast Star Advanced Power System （Suzhou） Co., Ltd.		 久之新管道机器人
Jiuzhixin Pipeline Robot		 随着我国社会经济及科学技术的飞速发展,在人们日常生活当中暖通 空调成为了不可忽视的重要的一个部分,为人们的生活带来了极大的便利, 同时给人们提供了舒适的环境.但是在暖通空调系统实际使用的过程当中 还存在诸多的问题,特别是管路系统等隐蔽工程的维护保养、故障检修等都 有诸多需解决的困难。 而随着社会的进步，机器人的使用越来越普及，管 道机器人可以携带传感器来进行完成管道内外作业的机电一体化系统。
技术需求： 1.检测管道各部位的风量； 2.检测管道局部风压； 3.检测管道的清洁度； 4.检测各主、次管道的风速； 5.可以对管道进行清洁和维护保养的工作； 6.对管道中检测出的问题或故障进行排除； 7.手机或电脑端可以和机器人连接、操作，并能提供实时数据至系统 中心。
注意事项： 1.因暖通管道的特殊性，圆形、方形等、机器人的结构设计需要考虑 适应性； 2.暖通系统应用于工业、商业等场所，一定程度上影响着人们的生活 和工作，需保障管道机器人的可靠性； 3.暖通管道环境较为复杂，在输送反馈信号等一系列信号时形成阻碍，有可能导致管道机器人停止工作。因此必须考虑管道机器人感应识别与越 障能力； 4.需考虑管道机器人控制系统自制完善自我修复功能。
With the rapid development of China's social economy and science and technology, HVAC has become an important part that cannot be ignored in people's daily life. It brings great convenience to people's lives and at the same time provides people with a comfortable environment. However, there are still many problems in the actual use of HVAC systems, especially the maintenance and troubleshooting of hidden projects such as piping systems, which have many difficulties to be solved. With the advancement of society, the use of robots has become more and more popular. Pipeline robots can carry sensors to perform mechatronics systems that complete the operations inside and outside the pipeline.
Technical requirements: 1. Detect the air volume of each part of the pipeline; 2. Detect the local wind pressure of the pipeline; 3. Detect the cleanliness of the pipeline; 4. Detect the wind speed of the main and secondary pipelines; 5. The pipeline can be cleaned and maintained Work; 6. Troubleshoot problems or failures detected in the pipeline; 7. The mobile phone or computer can be connected to and operated with the robot, and can provide real-time data to the system center.
Note: 1. Due to the particularity of HVAC pipes, the structural design of circular, square, etc. robots needs to consider adaptability; 2. HVAC systems are used in industrial, commercial and other places, which affect people’s lives and To work, it is necessary to ensure the reliability of the pipeline robot; 3. The HVAC pipeline environment is more complicated, which forms obstacles when a series of signals such as feedback signals are transmitted, which may cause the pipeline robot to stop working. Therefore, it is necessary to consider the pipeline robot's induction recognition and obstacle crossing ability; 4. It is necessary to consider the self-made and perfect self-repair function of the pipeline robot control system.
苏州亚科科技股份有限公司
Suzhou Yacoo Science Co., LTD		 开发一种新型环保可降解塑料
Develop a new type of environmentally friendly degradable plastic		 开发一种新型环保可降解塑料，该材料可光降解和生物降解，制备工 艺成本可控且有竞争优势，塑料材料具有优良的耐水、耐湿性能和良好的 力学性能，冲击强度高，承重性能好，使用范围时间长，并且具有较强可 降解性，降解率 100%。
Develop a new type of environmentally friendly degradable plastic, which is photodegradable and biodegradable, and the cost of the preparation process is controllable and has a competitive advantage. Plastic materials have excellent water resistance, moisture resistance and good mechanical properties, high impact strength, good load-bearing performance, and a long range of use. And it has strong degradability, with a degradation rate of 100%.
苏州亚科科技股份有限公司
Suzhou Yacoo Science Co., LTD		 开发一种脂肪胺合成共性氧化技术
Development of a common oxidation technology for aliphatic amine synthesis		 开发一种脂肪胺合成共性氧化技术，例如可将一甲胺氧化制备硝基甲 烷，乙胺氧化制备硝基乙烷，2-丙胺氧化制备 2-硝基丙烷，工艺环保，且 成本可控。
Develop a common oxidation technology for aliphatic amine synthesis. For example, monomethylamine can be oxidized to produce nitromethane, ethylamine can be oxidized to produce nitroethane, and 2-propanamine can be oxidized to produce 2-nitropropane. The process is environmentally friendly and the cost is controllable.
苏州亚科科技股份有限公司
Suzhou Yacoo Science Co., LTD		 开发一种高效环保的高纯双氟磺酰亚氨锂（LiFSI）的制备技术
Develop an efficient and environmentally friendly high-purity lithium bisfluorosulfonimide (LiFSI) preparation technology		 开发一种高效环保的高纯双氟磺酰亚氨锂（LiFSI）的制备技术，产能 纯度可达到 99.9%以上，痕量氯离子含量和金属杂质含量在 1-10ppm 以下， 水份控制在 100ppm 以内，原料成本控制在 200 元/kg 以内。
Develop an efficient and environmentally friendly high-purity lithium bisfluorosulfonimide (LiFSI) preparation technology, and the production purity can reach 99.9% or more. The content of trace chloride ions and metal impurities is below 1-10ppm, the moisture content is controlled within 100ppm, and the raw material cost is controlled within 200 yuan/kg.
苏州明俊仪器科技有限公司
Suzhou Mingjun Instrument Technology Co., Ltd.		 生物制药过程中流体传输用硅胶管原料配方开发
Formulation development of silicone tubing materials for fluid transfer in biopharmaceutical processes		 现生物制药用硅胶管的原料为 6 到 10 种硅胶粒子合成，主要供应的上 游是中石油中石化，但细化到制药级别的原料配方现只有外国 4 家供应商，分别为德国瓦克，美国杜邦，日本新月，还有一家美国的企业。 硅胶管是生物制药过程中的重要耗材，其中溶出、吸出、生物安全性 等 10 几项指标要求很高，国内现没有相关原料配方，所以此耗材如果外国 断货，将直接影响生物制药生产。 现硅胶管品牌有圣戈班（法国）、AP（美国）、新世界（美国）等国 外品牌，因国内销售价格很高，所以现市场需求量有 10 亿元人民币左右， 未来生物制大生产时，需求量将会大幅度提升。 通过排列组合实验，做出附合生物制药可以使用的硅胶管配方，打破 国外垄断。
The raw materials of silicone tubes for biopharmaceuticals are synthesized from 6 to 10 kinds of silica particles, and the main upstream supplier is China National Petroleum Corporation and Sinopec. However, there are currently only four foreign suppliers of raw material formulations refined to the pharmaceutical level, namely Wacker of Germany, DuPont of the United States, Crescent of Japan, and an American company. Silicone tubing is an important consumable in the biopharmaceutical process, among which 10 indicators such as dissolution, aspiration, and biological safety are very demanding. China does not currently have the relevant raw material formula, so if this consumable is out of stock in foreign countries, it will directly affect the production of biopharmaceuticals. The current silicone tube brands include Saint-Gobain (France), AP (United States), New World (United States) and other foreign brands. Due to the high domestic sales prices, the current market demand is about 1 billion yuan. In the future, when large-scale production of biological systems, the demand will greatly increase. Through permutation and combination experiments, a silicone tube formula compatible with biopharmaceuticals can be made, breaking the foreign monopoly.
迈杰转化医学研究（苏州）有限公司
MEDx (Suzhou) Translational Medical Co., Ltd.		 创新液体活检和诊断基因芯片
Innovative liquid biopsy and diagnostic gene chip		 合作开发高敏感性和高特异性肿瘤早筛早诊断和伴随诊断创新方法： 1）基于表观基因组学和/或蛋白质组学技术平台开发的用于肿瘤早筛早诊 的液体活检新方法；2）基于 DNA 芯片技术开发的用于肿瘤靶向治疗的伴随 诊断新方法。
Cooperate in the development of innovative methods for high-sensitivity and high-specificity tumor early screening and early diagnosis and companion diagnosis: 1) A new liquid biopsy method for early tumor screening and diagnosis based on epigenomics and/or proteomics technology platforms; 2) A new companion diagnostic method for tumor targeted therapy developed based on DNA chip technology.
苏州极目机器人科技有限公司
Suzhou Jimu Robot Technology Co., Ltd.(EAVison)		 开发基于多源信息大数据融合的作物-土壤信息一体化实时分析决策系统
Develop crops based on the fusion of multi-source information and big data Soil information integration real-time analysis and decision-making system		 通过对土壤肥力、墒情、作物养分、产量分布等多源数据的预处理分 析，构建面向作物-土壤信息一体化实时决策数据库。 研究基于大数据的土壤信息-作物养分的交互作用机制、多源异构数据 挖掘与融合技术以及作物-土壤信息一体化实时决策支持模型，针对不同主 栽作物，基于决策分析模型和专家知识库，实现实时施肥分析与决策。建 立基于植保数据集成云平台，通过实测土壤肥力-作物养分信息结合遥感数 据与历史积累数据，运用过程建模方法与深度学习技术，建立作物生长需 肥模型。 研究土壤肥力与作物养分交互作用机制，运用策略分析评价及 GIS 技 术等，开发基于大数据与多源信息融合的作物-土壤养分一体化实时分析决 策系统，建立农田变量作业 GIS 空间定位分析决策模型，开发基于多源信 息大数据融合的作物-土壤信息一体化实时分析决策系统，完成田间试验考 核与系统优化评估。
Through preprocessing and analysis of multi-source data such as soil fertility, moisture content, crop nutrients, yield distribution, etc., a real-time decision-making database for crop-soil information integration is constructed. Research on the interaction mechanism of soil information-crop nutrients based on big data, multi-source heterogeneous data mining and fusion technology, and crop-soil information integrated real-time decision support model. For different main crops, based on the decision analysis model and expert knowledge base, real-time fertilization analysis and decision-making can be realized. Establish an integrated cloud platform based on plant protection data, combining remote sensing data and historical accumulation data through measured soil fertility-crop nutrient information. Use process modeling methods and deep learning techniques to establish crop growth and fertilizer requirements models. Study the interaction mechanism of soil fertility and crop nutrients, use strategy analysis and evaluation and GIS technology. Develop a crop-soil nutrient integrated real-time analysis and decision-making system based on the fusion of big data and multi-source information, and establish a GIS spatial positioning analysis and decision-making model for farmland variable operations. Develop a crop-soil information integrated real-time analysis and decision-making system based on multi-source information and big data fusion, and complete field test assessment and system optimization assessment.
苏州真趣信息科技有限公司
Suzhou JoySuch Information Technology Co., Ltd.		 室内场景快速建模及在线可视化技术
Indoor scene rapid modeling and online visualization technology		 三维定位已经应用于多种行业。为快速可视化人员及设备整体情况， 需要快速高效建立所设计区域的城市三维模型，并实现基于浏览器的实时 动态可视化。目前城市或工厂三维模型复杂度较高，且不容易进行标准化 或批量化生成，但是由于定位展示的需求，对各个目标区域需要进行三维 建模，从而提高了项目的部署成本，特别是三维室内模型，精度要求高， 测绘成本大，为此需要研究快速三维模型生成技术，以及在线可视化技术， 从而提高项目部署与展示效果。
Three-dimensional positioning has been used in a variety of industries. In order to quickly visualize the overall situation of personnel and equipment, it is necessary to quickly and efficiently establish a three-dimensional urban model of the designed area. And realize the real-time dynamic visualization based on the browser. At present, the complexity of 3D models of cities or factories is relatively high, and it is not easy to standardize or generate batches. However, due to the needs of positioning and display, three-dimensional modeling is required for each target area, which increases the deployment cost of the project. Especially the three-dimensional indoor model requires high accuracy and high cost of surveying and mapping. For this reason, it is necessary to study rapid 3D model generation technology and online visualization technology to improve project deployment and display effects.
苏州芈图光电技术有限公司
Suzhou Mi Tu Optoelectronic Technology Co. Ltd. 		高速电芯片S 参数测试
High-speed electronic chip S parameter test		 高速电芯片 S 参数测试，芯片带宽 60G Hz，需要探针台，矢网（PNA）。
High-speed electronic chip S-parameter test, chip bandwidth 60G Hz, need probe station, vector network (PNA).