Mark Zuckerberg, CEO of Meta Platforms. Artificial intelligence has been an integral focus for the tech giant's leader amid competition from players like OpenAI, Microsoft and Google.Meta on Wednesday announced it's rolling out a new AI "world model" that can better understand the 3D environment and movements of physical objects.The tech giant, which owns popular social media apps Facebook and Instagram, said its new open-source AI model V-JEPA 2 can understand, predict and plan in the physical world. Known ...

i黑马 . 让创业者不再孤独@i黑马 以下文章来源于i黑马 ，作者i黑马 吴世春 梅花创投合伙人、黑马加速导师 早期创投，确实是比较难。 所以，如果以10年来看的话，我们肯定看过 40000-50000个 项目。 我一直相信： 所谓的收益，都是跟伴随着困难和高风险来的。 你能克服多少困难，你就能获得多少收益； 你能化解多大风险，你就能得到多少酬劳。 所以，早期投资肯定有它的难点，但是我相信， 只有极少数优秀机构能够坚守在早期。 现在一个企业从早期开始发展，要上 A 股、上科创板，大概需要 12-15 年。 所以，现在早期的钱，很多时候是支持不了这么长时间。 但我们也有一些新的办法，来去解决这样的问题。 我们一直坚守在早期投资领域，而且还能够就持续投资项目，说明我们有这样的能力。 现在我们一年大概要看 5000 个 项目，最后能够进入详细流程的大概是 1 000 个 项目，能 够上会的大概是 200 个项目，最后能过会的大概是 60 个项目。 目前，我们管理着有好几个城市的钱， 基金规模超100亿。 此前，我们投资、陪跑的企业已有 600多家 ，并陪伴 13家企业 完成了上市。 今年，我们至少还要 投五十家企业 ...

Representation of cryptocurrency and Circle logo displayed on a screen in the background are seen in ... More this illustration photo taken in Krakow, Poland on June 10, 2022. (Photo by Jakub Porzycki/NurPhoto via Getty Images)NurPhoto via Getty Images Circle Internet Group Inc (NYSE: CRCL), the issuer of the USDC stablecoin, has rapidly emerged as one of the market’s most watched stories following its IPO. Currently trading around $115 — a significant increase from its IPO reference price of $31 — the comp ...

Core Insights - Mainz Biomed is advancing its PancAlert project, focusing on a non-invasive blood test for early pancreatic cancer detection, aiming to confirm previous findings of 95% sensitivity and 98% specificity [1][5] Group 1: Project Development - The feasibility phase will verify candidate mRNA biomarkers and an accompanying machine learning algorithm in clinical blood samples to assess diagnostic performance [2][3] - The company plans to collaborate with Crown Bioscience, a CRO specializing in oncology, to support the verification process [2] Group 2: Objectives and Future Plans - The primary goal of the feasibility stage is to evaluate the robustness, reproducibility, and diagnostic performance of the assay, including sensitivity and specificity [3] - Following the feasibility phase, Mainz Biomed intends to conduct a validation study with a larger cohort, which is crucial for optimizing the test for clinical use and potential FDA submission [5] Group 3: Market Context - Early detection of pancreatic cancer is a significant unmet need in oncology, and the company is committed to improving patient outcomes through its diagnostic solutions [4] - The PancAlert project is part of Mainz Biomed's broader strategy to develop accessible molecular diagnostics for early cancer detection in areas with limited screening tools [5][7]

Core Viewpoint - Quest Diagnostics has announced a collaboration with The University of Texas MD Anderson Cancer Center to develop a blood test aimed at improving cancer risk assessment and screening for individuals at elevated risk [1][2]. Group 1: Collaboration and Test Development - Quest Diagnostics will develop and validate a laboratory-developed blood test based on circulating protein biomarkers linked to high cancer risk, including cancers such as colorectal, lung, breast, pancreatic, ovarian, liver, prostate, esophageal, and stomach [2]. - The test will utilize technology and intellectual property from the Multi-Cancer Stratification Test (MCaST), developed by Dr. Samir Hanash and his team at MD Anderson, which is based on extensive clinical research involving tens of thousands of individuals [2]. - If the test validation is successful, Quest may commercialize the test, aiming for availability to providers in North America by 2026 [3]. Group 2: Test Purpose and Market Need - The future test is intended to supplement conventional screening methods by providing insights that help identify patients who would benefit from appropriate cancer screenings, addressing the limitations of current methods that often target a limited number of cancers [4]. - Current screening methods can be invasive, inconvenient, or costly, leading to a significant number of patients skipping preventive screenings; only 51% of U.S. adults reported having a routine medical appointment or cancer screening in the last year [5][6]. - Quest aims to create a simple and affordable blood test to identify risks for various cancers, potentially encouraging patients to pursue preventive screenings and early medical assessments [6].

小微创投&投早投小 作为股权投资行业的从业者，之前写了不少本行观察的文章。从自身的从业来说，写这个题材，一、是为了吐槽一下干这行的不容易；二、是了解自己、 定位自身。我们这个行业的同行，大多时间是在研究被投项目的赛道分析，而很少去分析自己在行业位置定位和竞争力。圈外人觉得："私募"、"股 权"、"投资"、"天使投资人"......这些词汇，听起来都是相对比较高大上的名词。给别人投钱那是朝南坐，纯甲方！实际上，这个钱也不是大风刮来的，募 集就是求人，投完还是要退出。募集和退出，是纯乙方的活--求单和要账，这是全天下最难的几个事之一。真正做全整套链条，也只有自己知道：这活， 真不好干！ 在这三个前提下，我估计机构数量从10000多家注册，可能降到100家的规模了。系统中数量很多，但是活着的真不多了。而这样的一百家规模，其实也不 都是一个物种。 猫和虎，都是猫科动物，但完全不一样，也不是一回事。 中国股权行业从2000年初起步，到今天有快25年的历史了。机构数量最多时候有3万家注册，细分主业其实也是千奇百怪。如果把与我们类似的同类（比 如说猫科动物），可以先这样用这样的条件先筛一圈： 1. 有人民币基金，非纯美元； ...

RWA(Real World Asset,现实资产)项目的落地方式多种多 样，不同方案适用于不同的项目类型、目标市场和合规要求。以下是 五种主要的RWA 落地方案： 1)1.RWA 概念型落地方案 定义： 以宣传和教育为主，通过推广RWA概念、市场调研和初步试点，吸 引投资者和用户关注，为后续落地奠定基础。 A. 特点 · 主 要 以 概 念 验 证 (POC ) 和试点项目为主。 · 不涉及大规模的资产上链或代币发行。 · 注重市场教育和生态建设。 B. 优势 1. 成本低：不涉及大规模技术开发和合规投入，适合早期 探索阶段。 2. 灵活性强： 可根据市场反馈快速调整策略和方向。 3. 风险小：不涉及复杂的合规和资金流动问题，风险较低。 C. 劣势 1. 缺乏实际应用：仅限于概念验证，难以吸引实际投资者。 2. 市场认可度低：没有实质性资产支持，可能被视为炒作。 D. 适用场景 ●适合早期探索阶段，或希望通过教育和宣传建立市场认 知的项目。 2)2. 基于一级市场建设RWA 落地方案——实际资产 所有者发币 定义： 由实际资产所有者(如房地产、商品、债券等持有者)直接发行代币， 将现实资产上链，实现资产数字 ...

Core Insights - The article introduces the Reasoning as Meta-Learning (RaML) framework, which aims to reveal how large language models (LLMs) "think" by drawing parallels between reasoning and gradient descent optimization [1][2] - RaML posits that the reasoning trajectory generated by LLMs during problem-solving acts as a form of implicit parameter updates, leading to improved model performance [2][4] Group 1: RaML Framework and Mechanism - RaML's core insight is that the reasoning trajectory in LLMs resembles a "pseudo-gradient descent" process, where each reasoning step adjusts the model's internal state towards a better solution [2] - The framework decomposes the training process of LLMs into two levels: "inner-loop optimization" for specific tasks and "outer-loop optimization" for learning strategies across multiple tasks [8][9] - The study emphasizes that longer reasoning trajectories typically lead to better optimization outcomes, akin to more iterations in traditional optimization algorithms [14] Group 2: Empirical Validation and Performance - The QwQ-32B model's reasoning on the AIME24 dataset demonstrated that confidence in correct answers increases with the decoding of reasoning trajectories, supporting the idea of parameter updates through reasoning [3][4] - The comparison between supervised fine-tuning (SFT) and reinforcement learning (RL) models showed that SFT models outperform RL models in mathematical benchmarks, highlighting the benefits of guided learning [10][12] Group 3: Reflection Tokens and Optimization - The article discusses the role of "reflection" tokens in reasoning trajectories, which help the model reassess its outputs and improve performance by escaping local optima [15][17] - It contrasts "thinking" and "non-thinking" modes, indicating that forced early termination of reasoning can lead to suboptimal solutions, similar to premature stopping in gradient descent [18][20] Group 4: Generalization and Meta-Learning - The research indicates that LLMs trained on specific reasoning tasks can generalize to unseen tasks, leveraging learned universal features from various problems [21][23] - The RaML framework provides practical strategies for enhancing training performance by increasing the number of reasoning trajectories for each problem, akin to expanding the support set in meta-learning [25] Group 5: Future Directions and Efficiency - The article suggests exploring methods to extract shorter, equivalent optimization trajectories from longer reasoning paths to reduce decoding overhead while maintaining performance [27][30] - Initial experiments show that summarizing long reasoning trajectories can yield comparable results with significantly reduced computational costs, indicating a potential area for future research [30][31] Conclusion - The RaML framework offers a novel perspective on understanding LLM reasoning and training, revealing the intricate connections between reasoning, meta-learning, and gradient descent [32]

Baltic Horizon Fund (the Issuer) has initiated a written procedure to amend the terms and conditions of Baltic Horizon Fund EUR 42 million 5-year floating rate bonds maturing in 2028 (ISIN EE3300003235, the Bonds). The current outstanding nominal amount of the Bonds is EUR 18,999,997.80 as Baltic Horizon Fund has redeemed Bonds in nominal amount of EUR 23,000,002.20 in accordance with the terms and conditions of the Bonds (Terms and Conditions). The Issuer has decided to request consent from the holders of ...

Los Angeles, CA & Vancouver, BC, June 09, 2025 (GLOBE NEWSWIRE) -- BioSig Technologies, Inc. (Nasdaq: BSGM) ("BioSig”), which acquired Streamex Exchange Corporation as its wholly owned subsidiary (“Streamex”) in May 2025, is pleased to announce the appointment of Trevor Bacon and Kellan Grenier, co-founders of Parcl, as Strategic Advisors. Together, they bring deep expertise at the intersection of decentralized finance, synthetic asset creation, and real-world asset tokenization. Parcl is a leading blockcha ...