Ant Financial Releases Financial Large Model: Two Application Products Zhixiao Bao 2.0 and Zhixiao Zhu Will Launch After Filing

Ant Group unveiled a financial large model at the Bund Summit, based on its self-developed foundational large model, deeply customized for the financial industry, with a bottom-level computing cluster reaching a scale of ten thousand cards. The Ant Financial Large Model has demonstrated outstanding performance in financial specific tasks and has been comprehensively tested on wealth and insurance platforms. Ant Group also released two products based on the capabilities of the financial large model, namely the intelligent financial assistant "ZhiXiaoBao 2.0" and the intelligent business assistant "ZhiXiaoZhu 1.0", which will go live after completing the necessary filings. The Ant Financial Large Model is set to bring more natural interactions and more personalized service customization to the financial industry.

Shanghai Bank Launches Its First Hu-Shang Language Interactive AI Application to Support Smart Elderly Financial Services

Enterprises such as Shanghai Caiyue Star and Shanghai Bank signed a strategic cooperation agreement, launching the country's first complete Hu-Shang language interactive AI application, supporting elderly financial services and dialect intelligent system construction, providing more convenient financial services for elderly people who are accustomed to dialects.

Hang Seng Electronic Financial Model LightGPT Upgrade Launches Financial Intelligent Assistant Photon

The Hang Seng Financial Technology Conference announced the upgrade results and open product beta testing for the financial large model LightGPT. The LightGPT model has seen improvements in overall performance, safety compliance, and inference speed. Hang Seng introduced the financial intelligent assistant Photon to bridge the gap between large models and real-world applications. Intelligent products powered by Photon include Photon・Wenqu, Photon・Fangyuan, Photon・Shance, and Photon・Huiying. Hang Seng's subsidiary Hang Seng Juyuan also presented the latest product upgrades of the intelligent investment research platform WarrenQ.

OpenAI Claims ChatGPT's Gender Bias is Much Lower Than Our Own

OpenAI researchers recently discovered an interesting phenomenon: the usernames chosen by users when interacting with ChatGPT may subtly influence the AI's responses. However, this influence is generally very small and primarily limited to earlier, unoptimized versions of the model. This study delves into the differences in ChatGPT's reactions to identical questions when faced with usernames related to different cultural backgrounds, genders, and ethnicities. The choice of usernames as a focal point for the research was made because names often carry specific cultural, gender, and ethnic connotations.

Wenyuan Zhixing Releases Next-Generation Robotaxi GXR: Equipped with L4 Level Public Road Driving Capability

Wenyuan Zhixing officially launched its new generation mass-produced Robotaxi — GXR on October 15. GXR draws on more than 1,800 days of public operation experience in the Robotaxi field, featuring L4 level public road driving capability, and can handle complex scenarios such as morning and evening rush hours, mixed traffic, and high-speed driving at night. The Wenyuan Zhixing GXR is equipped with comprehensive redundancy systems for architecture, F-O line control chassis, computing units, and braking, effectively avoiding single point failures. It also features the new HPC 2.0 high-performance computing platform.

The Open Source Heygen is Here? TANGO Can Generate Full-Body Videos Matching Audio Indefinitely

TANGO, a powerful solution close to HeyGen. This innovative project not only supports facial and lip sync but impressively generates full-body action videos that perfectly match audio. The core advantage of TANGO lies in its unique generation logic. First, the system analyzes the short video samples provided by the user to construct an 'action map' containing various body poses and motion transitions. Then, it selects the best action sequences based on the input audio content. Finally, it creates seamless transition frames to produce a natural feel.