Bitget App
Trade smarter
Buy cryptoMarketsTradeFuturesEarnSquareMore

News

Stay up to date on the latest crypto trends with our expert, in-depth coverage.

banner
Flash
09:41
Opinion AI oracle made an error in judgment due to context contamination; the team will provide full compensation.
BlockBeats News, January 1st, the prediction market Opinion oracle was attacked. Due to "context contamination" during the open-source indexing phase of its AI oracle, a large amount of incorrect information was fed in, resulting in significant bias in prediction outcomes. The team has now provided the latest updates on the incident: 1. Full Compensation All affected users will receive full compensation. The team is currently collecting and verifying data, and compensation distribution will be completed as soon as possible. Details on the specific distribution progress will be announced separately. Affected users do not need to take any action—compensation will be credited automatically. 2. System Optimization (Completed) The factors that caused this context contamination have been successfully identified and fixed. This fix will further improve the accuracy and self-correction ability of AIOracle. 3. Additional Review Mechanism (Implemented) From now on, all market settlements will include an additional verification step before on-chain confirmation to ensure results are thoroughly reviewed and to prevent similar issues from occurring again.
09:40
Opinion AI Oracle Suffers Context Pollution Leading to Incorrect Judgment, Team to Fully Reimburse
BlockBeats News, January 1st: The Prediction Market Oracle Opinion was attacked due to a "Context Contamination" of its AI Oracle during the open-source indexing phase, where it was fed a large amount of misinformation, leading to a significant deviation in the prediction results. The team has now updated the latest developments of the event: 1. Full Compensation All affected users will receive full compensation. The team is currently conducting data aggregation and verification, and the compensation distribution will be completed as soon as possible. Specific details of the distribution schedule will be notified separately. Affected users do not need to take any action - the compensation will be automatically credited. 2. System Optimization (Completed) The factors that led to the context contamination in this incident have been successfully identified and rectified. This fix will further enhance the accuracy and self-correction capability of the AI Oracle. 3. Additional Audit Mechanism (Implemented) Subsequent market settlements will now include additional verification steps before on-chain confirmation to ensure that the results undergo comprehensive auditing and to prevent a recurrence of similar issues.
09:39
DeepSeek releases new paper by Liang Wenfeng: Proposes mHC new architecture to improve large model training stability
PANews, January 1st – According to Golden Ten Data, DeepSeek has released a new paper proposing a novel architecture called Manifold-Constrained Hyperconnection (mHC). This architecture aims to address issues in Hyperconnection Network (HC) technology, such as training instability and limited scalability caused by the disruption of the identity mapping property. The mHC architecture restores the identity mapping property by mapping the residual connection space of HC to a specific manifold, while also incorporating rigorous infrastructure optimization to ensure efficiency. This results in significant performance improvements and superior scalability. DeepSeek anticipates that mHC, as a flexible and practical extension of HC, will contribute to a deeper understanding of topological architecture design and point to promising directions for the evolution of foundational models. The paper lists Zhenda Xie, Yixuan Wei, and Huanqi Cao as co-first authors, with Wenfeng Liang also included among the authors.
News
© 2025 Bitget