人工智能算法

Search documents
科技赋能,米连科技助力青年寻爱
2 1 Shi Ji Jing Ji Bao Dao· 2025-10-11 10:49
30岁以下青年在婚恋社交中线上渠道占比上升,尤其是通过兴趣社群结识伴侣的比例增长,反映了青年 婚恋方式的多样化和线上化趋势。 这变化让青年们从被动等待,转向主动探索。另一位用户吉林姑娘余姚在伊对上,与任职政府机关的世 铭结缘,起初被彼此的城市、家境与对方稳定的工作吸引而奔现,一步步走到相知相恋。 当然,科技的介入并非一帆风顺。有用户就曾在微博上评论:"算法能算出兴趣契合,却算不出相处中 的包容。"她的疑问点出了潜在的隐忧,比如隐私被不慎泄露,或算法因数据偏差而偏向某些群体。 (原标题:科技赋能,米连科技助力青年寻爱) 某大厂95后程序员李林,最近从婚恋App伊对上结识了一位竹编爱好者小薇。两人从竹编技法聊到乡下 捉泥鳅,越聊越投缘感情也逐步升温,后来喜结良缘。 这并非什么天定缘分,而是科技悄悄架了座桥,让李林这类忙得脚不沾地、社交圈窄得可怜的年轻人, 让恋爱畅通便捷。 如今,广大青年面临工作节奏加快、社交圈受限的现实困境,婚恋难题越来越让人发愁。 李林的故事并非个例,它折射出科技如何悄然改变青年们的交友方式。过去,婚恋多靠亲友搭桥,局限 在熟人圈。如今,手机里的应用让一切变得触手可及。 作为北京米连科技旗下核 ...
这四类案件下月起由互联网法院集中管辖,最高法阐释调整意图
Xin Jing Bao· 2025-10-11 05:35
与旧规相比有何变化? 10月11日,《最高人民法院关于互联网法院案件管辖的规定》(以下简称《规定》)发布, 将"网络数 据权属、侵权、合同纠纷"等四类网络案件调整为由互联网法院集中管辖。《规定》自2025年11月1日起 施行。目前我国共有北京、杭州、广州三家互联网法院。 2018年9月,最高人民法院制发文件明确互联网法院审理案件范围,互联网法院集中管辖所在市的辖区 内应当由基层人民法院受理的网络购物、网络服务、网络金融借款、网络小额借款、网络著作权、网络 域名、网络侵权、网购产品责任、网络公益诉讼等十一类纠纷。 在此基础上,《规定》新增四类网络案件由互联网法院集中管辖。将 "网络数据权属、侵权、合同纠 纷"、"网络个人信息保护、隐私权纠纷"、"网络虚拟财产权属、侵权、合同纠纷"、"网络不正当竞争纠 纷"纳入互联网法院管辖范围。《规定》施行后,北京市、杭州市、广州市市辖区内应由基层法院审理 的上述案件将分别由三家互联网法院集中管辖。 此外,对于算法歧视、算法操纵、算法诱导沉迷、算法违法处理数据等现象,最高法研究室负责人认 为,此次调整有助于推动系统治理,从而促进平台经济、数字经济有序健康发展。 此次案件管辖调整 ...
网络数据权纠纷等案件将由互联网法院管辖,如何规范数据行为
Nan Fang Du Shi Bao· 2025-10-11 05:29
南都讯 记者刘嫚 发自北京 网络数据权属、侵权、合同纠纷等四类新型案件将纳入互联网法院集中管 辖。10月11日,最高人民法院发布《最高人民法院关于互联网法院案件管辖的规定》(以下简称《规 定》),对互联网法院案件管辖范围作出调整完善,新增四类新型网络案件由互联网法院集中管辖。 《规定》自2025年11月1日起施行。 最高人民法院研究室负责人对此表示,将网络数据权属、侵权、合同纠纷案件纳入互联网法院集中管 辖,有助于完善数据类纠纷裁判规则,明确数据权利义务边界、规范数据行为,统筹平衡数据权益保护 和数据交易流通效率,保障数据市场有序健康发展等。 互联网法院将集中管辖四类新类型案件 南都N视频记者了解到,2018年9月,最高人民法院制定印发《最高人民法院关于互联网法院审理案件 若干问题的规定》(法释〔2018〕16号,以下简称《2018年规定》),明确了互联网法院集中管辖所在 市的辖区内应当由基层人民法院受理的网络购物、网络服务、网络金融借款、网络小额借款、网络著作 权、网络域名、网络侵权、网购产品责任、网络公益诉讼等十一类纠纷。 最高人民法院研究室负责人在答记者问时对此调整解释,司法实践中,对一些新型、前沿、与 ...
马斯克新模型背后算法来自英伟达???
Sou Hu Cai Jing· 2025-09-26 00:19
Jay 发自 凹非寺 量子位 | 公众号 QbitAI Grok-4-fast最近在降本增效上的表现堪称一骑绝尘,连有「路由器」傍身的GPT5都给干趴下了。 面对如此惊艳的推理效率,很多人第一反应就是:堆卡式的算力scaling又一次大显神威了。 实际上,Grok的背后确实有英伟达的影子。 为此,英伟达研究团队推出了一种全新「混合结构」模型——Jet-Nemotron。 经过一系列全面的基准测试,发现Jet-Nemotron-2B的表现与Qwen3、Qwen2.5、Gemma3和Llama3.2等顶尖开源模型不相上下,还能实现约53倍的速度 提升。 例如在MMLU-Pro上,Jet-Nemotron-2B不仅准确率比Qwen3-1.7B-Base更高,就连生成速度也要快上47倍。 此外,Jet-Nemotron-2B即便遇上参数更大的模型也丝毫不虚,它在MMLU和MMLU-Pro上的准确率甚至可以超过DeepSeek-V3-Small和Moonlight(总参数 量15B,激活参数量2.2B)。 但这次立功的,或许不是老黄的显卡,而是算法。 没错,Grok-4-fast的秘密武器,被和一篇英伟达的算法论文关联 ...
油菜高准确度基因组预测新模型将品种培育周期缩短至4年
Ke Ji Ri Bao· 2025-09-19 07:23
Core Insights - The research team from the Chinese Academy of Agricultural Sciences has developed a high-accuracy genomic prediction model for rapeseed, integrating 404 global rapeseed resources, which provides efficient "precise navigation" for target breeding related to flowering time, oil content, and yield [1][2] - Traditional breeding methods for rapeseed take up to 6 years, which is insufficient to meet the urgent demand for high-quality varieties in agricultural production [1] - The new model has achieved over 90% prediction accuracy for flowering time and thousand-grain weight, and over 83% for oil content and pod length, significantly enhancing the prediction accuracy for complex traits [1] Industry Impact - The integration of genomic data and artificial intelligence algorithms with the genomic prediction model has led to a rapid breeding technique, reducing the breeding cycle from 6 years to 4 years [2] - The varieties bred using this model can establish a unique "gene ID," enabling full-process gene traceability from breeding to production, thus providing a foundational guarantee for the high-quality development of the rapeseed industry [2]
中原金太阳申请考虑碳捕捉效益的配电网内风电容量区间计算方法专利,实现碳效益‑经济成本的动态权衡
Jin Rong Jie· 2025-08-23 01:21
Group 1 - The company Henan Zhongyuan Jinyang Technology Co., Ltd. has applied for a patent titled "A Calculation Method for Wind Power Capacity Range in Distribution Networks Considering Carbon Capture Benefits" [1] - The patent application was published under CN120524785A and was filed on March 2025 [1] - The invention relates to the field of wind power capacity configuration and involves a method that integrates artificial intelligence algorithms with physical laws of energy systems [1] Group 2 - Henan Zhongyuan Jinyang Technology Co., Ltd. was established in 2020 and is located in Zhengzhou, primarily engaged in technology promotion and application services [2] - The company has a registered capital of 90 million RMB and has invested in 41 enterprises [2] - The company has participated in 91 bidding projects and holds 21 patents, along with 6 administrative licenses [2]
物流网络货运平台的创新力量:跨越速运的智慧物流之路
Sou Hu Cai Jing· 2025-08-20 02:19
Core Insights - The rise of logistics network freight platforms marks a new era of intelligence in the logistics industry, with KuaYue Express emerging as a key player driving industry development through technological strength and innovation [1][9] Group 1: Intelligent Dispatch - KuaYue Express's logistics network freight platform features an intelligent dispatch system that utilizes advanced big data analysis and AI algorithms to match vehicles and cargo in real-time, significantly reducing transit time and transportation costs [3][9] Group 2: Digital Services - The platform incorporates IoT and blockchain technologies to enhance transparency, safety, and convenience in logistics services, allowing customers to monitor cargo status in real-time, including critical information such as location, temperature, and humidity [4][9] Group 3: Personalized Services - KuaYue Express offers customized logistics solutions to meet diverse customer needs, whether for large-scale e-commerce shipments or small parts delivery in manufacturing, thereby increasing customer satisfaction and loyalty [6][9] Group 4: Continuous Innovation - The company's success is attributed to its ongoing commitment to innovation, investing in R&D to explore new technologies that enhance logistics service efficiency and quality, keeping KuaYue Express at the forefront of industry innovation [7][9]
Haverty Furniture(HVT) - 2025 Q2 - Earnings Call Transcript
2025-07-31 15:00
Financial Data and Key Metrics Changes - Company reported Q2 2025 sales of $181 million, a 1.3% increase year-over-year, with comparable store sales down 2.3% [3][18] - Gross margin improved to 60.8% from 60.4%, reflecting product selection and merchandising mix [18] - Pre-tax profits decreased to $4.3 million, with an operating margin of 2.4%, down from $6.5 million and 3.6% in Q2 2024 [3][18] - Earnings per share (EPS) for the quarter was $0.16, compared to $0.27 in the same quarter last year [3][18] Business Line Data and Key Metrics Changes - Total written sales increased by 0.4%, while design and special order business saw a mid-single-digit decline due to tariff impacts [4][9] - Average ticket size decreased slightly to just under $3,400, while designer average ticket grew approximately 5% to over $7,600 [4][9] - Upholstery and bedroom categories outperformed with low to mid-single-digit positive sales, while dining room and decor categories experienced high single-digit declines [9] Market Data and Key Metrics Changes - Traffic remained positive in the mid-single digits compared to the same period last year, with a notable increase during the Memorial Day event [4][6] - Organic traffic increased by 15.6% following the implementation of Adobe's Edge delivery service [7] - Web sales grew by 8.4% for the quarter, attributed to improved digital marketing strategies [7] Company Strategy and Development Direction - Company aims to return to positive same-store sales and is focused on enhancing customer experience through new point of purchase and tagging programs [10][12] - Plans to open five new stores annually, with two new stores in Houston and one relocation in Daytona Beach planned for 2025 [12][14] - Company is actively managing supply chain challenges and tariff uncertainties while maintaining gross margin guidance [11][15] Management's Comments on Operating Environment and Future Outlook - Management noted a struggling housing market with high interest rates and inflation concerns but highlighted consumer resilience [4][15] - Confidence in maintaining gross margin guidance despite potential tariff impacts, with proactive vendor communication [11][22] - Management expressed optimism about gradual improvement in sales trends and plans to invest more in marketing strategies [37][48] Other Important Information - Selling, general, and administrative expenses increased by 4.1% to $107.3 million, representing 59.3% of sales [19] - Company has no funded debt and ended the quarter with $107.4 million in cash and cash equivalents [20][21] - Anticipated capital expenditures for 2025 remain at $24 million, focusing on new store openings and IT investments [23] Q&A Session Summary Question: Can you speak to the cadence of your written sales throughout the quarter and any notable regional differences? - Written business was down around 2% in April, up almost 1% in May, and up around 2.5% in June, with no significant regional differences noted [26] Question: Can you quantify the impact of suspending special orders from China on same-store business? - Management acknowledged the impact on design business but could not quantify the exact effect [27][28] Question: Have you taken any pricing actions regarding tariffs, and what are your expectations? - Pricing adjustments were made in May, and management is prepared to adjust pricing based on final tariff outcomes [29][30] Question: What marketing strategies do you believe will be most impactful in driving same-store sales? - New pricing strategy and successful marketing campaigns, including extended promotions, are expected to drive traffic and conversion rates [35][36] Question: How do you view the promotional environment across the industry? - Management feels confident in their promotional strategies and plans to increase marketing investments while maintaining brand integrity [40][42] Question: What is the outlook for store openings and the real estate environment? - Store openings have been pushed to 2026, but management remains optimistic about finding suitable locations and maintaining reasonable rents [49][51]
科学与健康|健康“链”世界 中国医疗装备走出国际范儿
Xin Hua She· 2025-07-25 14:39
Core Insights - China's medical equipment industry is rapidly emerging, with exports covering over 9,000 medical institutions in more than 190 countries and regions by 2024 [2] - The industry has seen significant growth, with revenue from large-scale medical equipment enterprises exceeding 540 billion yuan, maintaining a rapid growth trend over the past decade [5] - The approval of innovative medical devices has accelerated, with 45 new devices approved in the first half of the year, marking an 87% year-on-year increase [5] Group 1: Industry Growth and Innovation - The medical equipment market in China has reached a scale of 1.35 trillion yuan, showcasing the country's position as a major application market and production base for medical equipment [5] - Innovative products such as smart bionic hands and advanced imaging systems are being developed, demonstrating the integration of brain-machine interface technology and artificial intelligence [5][9] - The international competitiveness of China's medical equipment industry has been improving, with a compound annual growth rate of 9.4% in import and export trade over the past five years [8] Group 2: Global Expansion and Collaboration - Chinese medical equipment is increasingly being utilized globally, with products like CT scanners and ventilators being exported to serve numerous patients [8] - The shift from single equipment exports to comprehensive professional solutions is evident, as companies like Mindray Medical have developed integrated ecosystems to enhance operational efficiency [8][10] - Collaborative efforts are being made to strengthen the ecosystem of production, education, research, and application, with initiatives from various government bodies to support innovation in the medical field [10][11]
控价公司如何快速识别低价商品?
Sou Hu Cai Jing· 2025-07-23 02:52
Core Viewpoint - The article outlines methods employed by price control companies to quickly identify low-priced products across various e-commerce platforms, ensuring compliance with pricing regulations and maintaining brand integrity. Group 1: Technical Monitoring - Deployment of intelligent monitoring systems, such as Dingdian Data, utilizing self-developed AI crawling systems to scan over 10 major e-commerce platforms daily for price monitoring [3] - Setting price alert thresholds, with alerts triggered when prices fall below 85% of the official price, including a three-tier warning system for different price drop levels [3] - Use of image recognition technology to automatically compare product images with official galleries, identifying hidden violations like price tampering [3] - Comprehensive monitoring across various platforms, focusing on low-price keywords, SKU price fluctuations, and compliance of promotional language [3] - Price gap scanning technology to automatically capture store prices and compare them with official pricing, highlighting links that fall below alert thresholds [3] Group 2: Manual Review - Manual review of promotional activities, such as live streaming sales and bundled offers, to avoid technical misjudgments [4] - Monitoring of keywords commonly associated with low-price promotions, such as "special price," "clearance," and "discount" [4] Group 3: Data Processing and Analysis - Generation of a list of top 10 violating stores, indicating the extent of price reductions, platform distribution, and historical violation records [5] - Establishment of a database for violating stores, documenting historical violations, store qualifications, and logistics information to assess potential cross-selling [5] - Utilization of professional tools to create visual reports displaying the distribution of low-price links, information on violating merchants, and price fluctuation trends [5] Group 4: Additional Methods - Verification of sales stores' authorization to sell brand products, as unauthorized sellers often attract consumers with lower prices [7] - Quality checks on low-priced products, including packaging, labeling, and production dates, to determine authenticity [7] - Batch number tracing to compare legitimate product batch numbers with those on suspicious links, facilitating the identification of infringement [7]