Zhu Guang, CEO of Du Xiaoman: Actively deploy the application of large model technology in the financial field

On March 9 th, Zhu Guang, CEO of Du Xiaoman, said in a signed article that Du Xiaoman has been actively deployed in exploring the application of large model technology in the financial industry.

Large models can be divided into comprehensible large models and generative large models. In Zhu Guang’s view, the comprehensible big model raises the decision-making ability of customer management and risk control to a new height and reduces financial risks; Generative AI will focus on customers, reshape the customer service process and experience, and become an "always-on super financial consultant" and an "all-round business assistant". This means that big model technology is defining a new entrance to the industry.

It is reported that Du Xiaoman has developed a series of applications based on large models relying on Baidu artificial intelligence technology. Taking risk management as an example, Du Xiaoman has applied LLM, a large language model, to the interpretation of Internet text data and credit report. Through the pre-training model constructed with text data and AI algorithm, he can interpret the credit report into 400,000-dimensional risk variables and better identify the credit risk of small and micro business owners. With the iteration of the model, the potential of the large model in intelligent risk control will be further released.

The self-developed voice robot can realize multiple rounds of complex conversations, with a non-inductive rate of 99%, that is, 99% of users don’t know that it is actually a robot talking to themselves, which can replace more than 60% repetitive work in credit business, and is widely used in marketing and collection work of many financial institutions, which greatly reduces labor costs and also has certain emotional recognition ability.

ChatGPT’s semantic understanding ability is also the key investment direction of Du Xiaoman in the field of natural language processing (NLP). In 2021, in the document sorting task of MS MARCO competition held by Microsoft, the NLP team ranked first and set a new record; The XuanYuan pre-training model developed by the team also ranks first in the CLUE classification task.

Baidu’s generative dialogue product "Wen Xin Yi Yan" (English name: ERNIE Bot) based on Wen Xin big model technology opened up ecological cooperation, and Du Xiaoman became the first financial technology company to access it. Next, based on the large-scale model technology base of Wenxin, combined with the financial industry knowledge and data accumulated in Xiaoman’s business scenario, Xiaoman will create a brand-new intelligent customer service, intelligent risk control and intelligent interactive service.

Zhu Guang said that the big model technology like GPT is the core topic of the next generation AI technology contest, which will have a profound impact on the intelligence level and digitization level of the financial industry. Financial technology companies should seize the opportunity window of this technological change, actively explore the application and development of large-scale model technology, and help the financial industry realize the leap from digitalization to intelligence.

Nature writes: In the next 30 years, cancer will cost the world 25 trillion US dollars.

Monoclonal antibody against B-cell psoriasis leukemia, a target of cancer artificial intelligence

①Nature:Cancer will cost the world $25 trillion over next 30 years

In the next 30 years, cancer may cost the world $25 trillion.

DOI :10.1038/d41586-023-00634-9

②Seminars in Cancer Biology:Artificial intelligence in cancer immunotherapy: applications in neoantigen recognition, antibody design and immunotherapy response prediction

Review: Application and development of artificial intelligence in the field of precision oncology: new antigen recognition, antibody design and immunotherapy response prediction.

DOI :10.1016/j.semcancer.2023.02.007

③Molecular Cancer:CD39/CD73/A2AR pathway and cancer immunotherapy

Review: New progress in pathophysiological function of CD39/CD73/A2AR pathway in cancer development, metastasis and drug resistance.

DOI :10.1186/s12943-023-01733-x

④Immunity:B cell receptor signaling in germinal centers prolongs survival and primes B cells for selection

B cell receptor signal transduction in germinal center prolongs the survival of B cells and initiates the selection of B cells.

DOI :10.1016/j.immuni.2023.02.003

⑤Frontiers in Immunology:Metabolic rewiring in keratinocytes by miR-31-5p identifies therapeutic intervention for psoriasis

The metabolic recombination of miR-31-5p in keratinocytes determines the intervention on psoriasis.

DOI :10.15252/emmm.202215674

⑥Journal of Hematology & Oncology:Novel agents and regimens in acute myeloid leukemia: latest updates from 2022 ASH Annual Meeting

The latest drug research and new scheme progress of American Society of Hematology (ASH) for acute myeloid leukemia (AML)
DOI :10.1186/s13045-023-01411-x

⑦FDA extended the review of adult Roctavian with severe hemophilia A because of "significant changes in a large number of additional data"

⑧ A small molecule MET inhibitor, gumetinib, which was developed by Shanghai Institute of Pharmacology, China Academy of Sciences, was approved for marketing, and it was used for locally advanced or metastatic non-small cell lung cancer (NSCLC) with the jump mutation (METex14 jump) in exon 14 of mesenchymal-epithelial transforming factor (MET).

⑨ The clinical study of Relma-cel targeting CD19 for first-line high-risk large B-cell lymphoma was officially launched.

⑩ Lilly will terminate the research and development of Solanezumab for Alzheimer’s disease because it fails to slow down the cognitive decline of preclinical Alzheimer’s patients.

The world of scientific exploration is full of challenges. AtaGenix pays tribute to every pioneer and is more committed to becoming your trusted partner!