
可信光语大模型是由无限光年(INF)公司在2024年世界人工smart (phone, system, bomb etc)Conference (WAIC), an innovative big model that combines big language modeling with symbolic reasoning.
Background and release
- Company Background: Infinite Lightyear is an innovative company focusing on trusted big model technology. Its founding team has rich experience in large-scale industrialization of the Internet and industry expertise, and most of its members come from well-known enterprises and financial institutions such as Alibaba, Ant Group, ByteDance, Microsoft, and so on.
- Release time: On July 4, 2024, at the World Conference on Artificial Intelligence 2024, Infinite Light Years released the Trusted Light Language Grand Model and its technical report.
Technical characteristics
- Combining Big Language Modeling and Symbolic Reasoning::
- The Trustworthy Optical Language Grand Model not only possesses the extensive knowledge and language generation capabilities of a grand language model, but also incorporates symbolic reasoning techniques, thus effectively solving the illusory problem prevalent in grand models and enhancing the credibility and reliability of the model.
- Improving Inference Accuracy and Reducing Service Costs::
- As a "specialized" tens of billions of big models, Trusted Optical Language Big Model has performed well in financial and medical verticals, surpassing OpenAI's trillion-parameter big model, GPT4-Turbo, and its efficient inference and accurate knowledge fusion not only improves inference accuracy, but also reduces service costs by optimizing algorithms and Its efficient inference capability and accurate knowledge fusion ability not only improve inference accuracy, but also reduce service cost by optimizing algorithms and architectural design.
- semiotic technique::
- Infinite Lightyear uses neural symbolic technology to organically combine the reasoning ability of the symbolic school with the learning ability of the associative school, which dramatically improves the trustworthy generation ability of generative AI in industry applications, reduces the error rate, and improves reliability. This technological route allows large models to improve the accuracy of reasoning in application scenarios while reducing the cost of services.
Industry Applications and Achievements
- Financial sector::
- Based on the trusted big model technology, Infinite Lightyear has launched an AI investment research assistant in cooperation with head securities companies. This investment research assistant is the first in China to realize accurate, efficient and in-depth interpretation of corporate financial reports with big models, and provide analysts and other professionals with one-stop functions including information summary and analysis, paradigm writing and accurate search. Currently, the product has been used by a number of securities companies and asset management organizations.
- Medical field::
- Infinite Lightyear cooperates with domestic head tertiary hospitals to jointly build a big model of the medical industry based on trusted technological innovation. The big model effectively assists doctors in improving the efficiency and accuracy of reports in the interpretation of medical examination reports, providing strong support for the digital transformation of the medical industry. At the same time, Infinite Lightyear's medical big model has excellent performance on the authoritative medical data set MedBench, and is the first medical big model with more than 90 points on both lists, with a comprehensive score of 90.4 on the API list, and ranked the first in five dimensions, including complex medical reasoning.
future outlook
Infinite Lightyear team will continue to explore the new generation of credible "gray box" big model technology, so that the vertical big model is more accurate and credible, and become a "financial advisor", "professional doctor" and more experts in the field. They are committed to promoting the use of AI technology in various fields. They are committed to promoting the in-depth application of AI technology in various industries, so that AI technology can be used by everyone and benefit all walks of life. At the same time, Infinity Lightyear will continue to pay attention to the latest developments and technology trends in the field of artificial intelligence, and continue to introduce more advanced and practical AI products and solutions to contribute to the development and progress of society.
summarize
Trustworthy Optical Language Big Model is an important innovation of Infinite Lightyear in the field of artificial intelligence, which combines Big Language Model with symbolic reasoning technology, effectively solves the illusion problem of Big Model, and improves the trustworthiness and reliability of the model.
data statistics
Relevant Navigation

ByteDance launched a self-developed big model. Through byte jumping internal 50 + business scene practice verification, daily 100 billion tokens large use of continuous polishing, to provide multi-modal capabilities, with high quality model effect for the enterprise to create a rich business experience

SKYMEDIA
Wanxing Technology has developed China's first audio and video multimedia creation pendant big model, which integrates video, audio, picture and language processing capabilities to provide powerful AI creation support for the digital creative field.

TianGong LM
Kunlun World Wide's self-developed double-gigabyte large language model, with powerful text generation and comprehension capabilities and support for multimodal interaction, is an important innovation in the field of Chinese AI.

QwQ-32B
Alibaba released a high-performance inference model with 32 billion parameters that excels in mathematics and programming for a wide range of application scenarios.

HunyuanImage2.1
Tencent launched the open source raw image model, which natively supports 2K HD raw images, accurately parses complex semantics, and can efficiently generate high-quality images with Chinese and English fusion.

Seedream 2.0
Byte Jump launched a native bilingual image generation model with excellent comprehension and rendering capabilities for a wide range of creative design scenarios.

WebLI-100B
Google DeepMind launches a 100 billion visual language dataset designed to enhance the cultural diversity and multilingualism of AI models.

s1
An AI model developed by Fei-Fei Li's team that achieves superior inference performance at a very low training cost.
No comments...
