What is ShiShen
ShiShen is the first vertical large model in the cooking domain launched by Robam Appliances. Based on the massive cooking data accumulated by Robam Appliances over 46 years and combined with DeepSeek technology, it can accurately understand the needs of cooking scenarios. Users can interact with "ShiShen" through voice, text, and other methods to obtain services such as personalized recipe recommendations and healthy meal planning. It can integrate with Robam Appliances' digital kitchen appliance products to realize automation and intelligence in the cooking process. ShiShen can explain cooking cultural stories, bringing users emotional companionship and cultural experiences.
Main Functions of ShiShen
- Personalized Recipe Recommendations: ShiShen can accurately recommend recipes with different flavors to users based on their dietary preferences, health data, historical cooking records, and health status information obtained through facial recognition.
- Healthy Meal Planning: By analyzing users' medical examination reports and health status, ShiShen can generate personalized healthy meal plans for users.
- Multi-Device Integration: It seamlessly integrates with Robam Appliances' digital kitchen appliance products, such as range hoods, stoves, steam ovens, etc., to realize automation and intelligence in the cooking process.
- Emotional Companionship and Cultural Experience: ShiShen provides assistance with cooking functions, explains the cultural stories and nutritional knowledge behind cooking, and recommends dishes with cultural characteristics according to different festivals or occasions.
- Multimodal Interaction and Perception: The ShiShen large model supports multimodal interaction, including multiple input methods such as voice, text, and images. It can perceive users' health status through facial recognition technology and, combined with users' dietary records, medical examination reports, and other information, provide personalized meal planning for users. Through functions such as real-time announcements by smart cooking speakers and automatic adjustment of stove heat, it can integrate with kitchen appliance devices.
Technical Principles of ShiShen
- Accumulation and Integration of Massive Data: The ShiShen large model integrates the Chinese cooking knowledge base accumulated by Robam Appliances over 46 years, privately developed recipe data, and real user behavior data. The data covers multidimensional information such as cooking culture, nutrition, and health, forming its data moat. It combines 2TB of public-domain and private-domain knowledge bases, and through analysis and learning of massive data, it achieves a deep understanding of cooking scenarios.
- Fine-Tuning and Optimization Based on Large Models: The ShiShen large model is based on the open-source InternLM dialogue model from Shanghai AI Laboratory, and uses Xtuner for LoRA fine-tuning on the XiaChuFang Recipe Corpus dataset of more than 1.52 million Chinese recipes. In this way, the model can better adapt to the specific needs of the cooking field and generate personalized recipes and cooking suggestions.
- Scenario-Based and Full-Chain Solutions: The ShiShen large model can realize full-chain solutions from recipe recommendations and ingredient preparation before cooking, to device control and heat adjustment during cooking, and then to cleaning equipment after cooking. It can generate recipes according to users' needs and, through collaboration with Robam Appliances' full-chain products, complete complex cooking tasks.
- Continuous Learning and Optimization: The ShiShen large model has the ability to continuously learn and can continuously optimize algorithms based on user feedback and new data. The dynamic learning mechanism can better adapt to users' personalized needs and improve the cooking experience.
Application Scenarios of ShiShen
- Home Cooking: Users can quickly obtain suitable recipes through "ShiShen" and integrate with kitchen appliance devices to complete cooking.
- Health Management: Provide users with healthy meal planning to help them maintain healthy eating habits.
- Cultural Experience: Enhance users' understanding of and interest in culinary culture by explaining culinary cultural stories and recommending specialty dishes.