July 2009 No. 1 percentage point company was born in a small apartment in Zhongguancun. At this point, the tide began to rise coincided with big data. Is called its Xing Yan Bo, Cloudera, HortonWorks, MapR data and other large companies sought after by the market, a stand at the forefront, thunder Missing Piece beach-goers.
Baiju too Jian, a flash ten years have passed. Ebb Tide of the market, companies have to make big data survival of the fittest. Surviving company, experienced a market test and live baptism; and some of the former stars of big data, but a difficult one, people can not help but lament its decline sudden fall. Is undoubtedly one percentage point has withstood the test of the market and the baptism of the company, in its ten years of development history, not only witnessed the Chinese big data industry from scratch, from the rise of the whole tortuous process towards landing; itself gradually explore a their own road of development, the first to achieve profitability in China big data companies, the value of the data on the road floor walk in the forefront of the industry.
"The company set up its mind is to use data intelligence technology to provide users continue to create measurable value. Since this decade, the percentage is always adhere to the original intention." Chairman and CEO percent Meng Su says.
Baptism decade, exploring the road of landing data value
Big data companies the hardest is that: the value of the data of previous landing this thing is not done, you need to continue to explore and practice, even if it is valued as high as $ 40 billion in the United States big star Palantir companies are no exception. This requires big data companies not only need to master the most advanced data intelligence technology, also need to realize the integration of technology with industry scene, let the data applications and smart technology truly play value in the business scene.
Percent chairman and CEO Su Meng
We can say that the first decade of exploration data value is the percentage of landing decade. In percentage chairman and CEO Su Meng view, percentage points over the past decade has experienced three important transformation, gradually embark upon a path of landing a data value.
The first transformation occurred in 2014, points to provide personalized recommendation from the pure SaaS service for the Internet industry, transformed into providing enterprise-class services. This transformation means that percentage of the customer base expanded from the Internet industry to more traditional industries. Points previously accumulated in the service of many electricity providers and network media rich massive data processing and analysis capabilities precipitation, have launched a large data operating system (BD-OS), industry AI decision-making system DeepMatrix other data intelligence products, promote application data intelligence technology and products in traditional industries.
The second round of restructuring is to enter the overseas market, percentage points in 2016, prospective overseas markets saw a huge demand for space for big data and artificial intelligence, and the first to respond, "along the way" initiative, became the first layout of the overseas business of big data company . Su Meng view, overseas business is a typical "strong customer" provides a rare "strong scene," the country needs a lot of big data and artificial intelligence to complete the digital transition as well as to enhance the governance level, these requirements to the construction of state-level points big data and opportunities wisdom of the brain, so that the whole stack large percentage of data technology products and intelligent recognition technology products has been fully applied in overseas more than ten countries and has research from requirements to process standards and service capabilities to deliver implementation.
The third round of restructuring is to enter the government sector, especially the "Digital China" to speed up construction, deep plowing percentage points in the last two years the government level market, including digital government, such as provincial or municipal government big data platform; the other is public safety field, which has created DeepFinder points and other more mature solution, and has been in-depth application on multiple domestic users.
Three percentage points from the transformation, we appear to feel a great percentage of "cross-border" range, but the essence has not changed, Su Meng said that these points are in the dry years one thing, the data is the use of intelligent technology to help various industries realize the value of data floor. "It's ten years along the way, the process of transformation in the harvest three times a lesson, but the most important thing is to find the right direction, but also the formation of a large data set, including full-stack technology products, cognitive intelligence engine, intelligent decision-making applications including mature product system for the future lay the foundation for sustainable development. "
Focus cognitive intelligence, exploration technology and integration scenarios
With the rise of 5G, networking, artificial intelligence, big data technology, a new technology gushing and hair era is taking shape, and the convergence between the fusion between technology and art scene is also becoming increasingly apparent. For points, the past decade has been committed to intelligent data associated with innovation, and walk in the forefront of technology floor. In the future, the focus points will focus on intelligence, knowledge and technology and the depth of integration scenarios.
百分点CTO刘译璟表示，从纯技术的角度来看，未来三到五年包括人工智能在内的多项技术在应用场景上主要体现三个趋势：首先是实时性，尤其是5G、IoT等技术发展起来之后，需要系统能够短时间做出判断和决策；第二则是交互性，人机交互的场景会越来越多，这种场景将会给技术带来诸多挑战；第三则是自主性，很多系统和设备逐渐不需要人的监管，可以自主做出决策。
百分点CTO刘译璟
正因为如此，百分点在产品和技术上也紧跟趋势。首先，百分点的数据中台产品已经在多个行业得到很好的应用，拥有完善和健全的产品、工具集以及行业实践，为进一步提升数据处理和分析的实时性，百分点加快了数据中台产品与AI技术的融合，让数据标注、数据治理、脚本生成等更加快速和自动化，并且对上层业务进行一些AI层面的支持。“未来，百万亿级的数据处理规模可能非常常见，这就需要我们在技术层面对实时性进行更加有效的保障。”
对于越来越多的人机交互场景，百分点也有诸多探索，并且推出了Clever BI这种引领未来趋势的商业智能产品，用户以自然语言交互的方式与系统进行可视化交互，获取和分析数据，大幅降低了数据分析的门槛。不仅如此，Clever BI还具备完善的数据管理方案和强大的增强型分析，可以支撑没有专业背景的用户完成数据的“数据-准备-整合和分析”的全过程。
认知智能技术会是系统自主决策突破的关键。认知智能技术也是百分点未来重点希望突破的领域。百分点首席算法科学家苏海波介绍，百分点去年成立了认知智能实验室，重点聚焦在自然语言处理、知识图谱、智能交互等技术上。“机器在人脸识别等感知层的准确率已经超过人类。但是在认知智能层面，还需要更长的时间来取得突破。”
在苏海波看来，人工智能技术现在融合的趋势非常明显，认知智能技术突破的关键点，在于如何将连接主义、符号主义和行为主义三个流派进行深度融合。百分点对于认知智能主要进行三个方向的重点研究：
百分点公司首席算法科学家苏海波
首先是深度学习与迁移学习的融合，深度学习依赖大量的标注数据，但是现实情况却是数据样本严重不足，通过迁移学习将知识迁移到新的领域，可以降低所需要的数据样本。百分点在海外多个案例中就遇到过样本匮乏的情况，通过深度学习与迁移学习的融合，很好地解决了这些问题。
其次就是深度学习与知识图谱的融合，深度学习属于人工智能连接主义流派，而知识图谱则是符号主义流派。知识图谱是知识的表示，通过与深度学习融合起来，可以得到更好的自然语言理解模型，然后再反馈到知识图谱中，从而发挥到更大的作用。百分点在公共安全领域的多项应用与实践就是这两种技术融合的代表。
第三是深度学习与强化学习的融合。强化学习是人工智能行为主义的代表，深度学习与强化学习融合是连接主义和行为主义的融合，可以让系统不断从自身经验中学习，让系统的一系列决策做出更好更鲁棒的决策，被认为对未来人工智能会产生重要的影响。
据悉，百分点公司认知智能实验室已经在深度学习+迁移学习、深度学习+知识图谱、深度学习+强化学习等方面在多个行业应用中进行了大量的探索与实践。未来，百分点公司的认知智能战略会以“强场景”为驱动，将NLP、知识图谱、智能交互等认知智能技术，应用到企业的市场化决策和政府的公共安全决策两个方向。
未来：开启数据智能黄金十年
十年弹指一挥间。
百分点已经从当年的纯SaaS公司成长为数据智能领域的头部企业，得到了市场的广泛认可。在苏萌看来，百分点的初心未变，希望利用数据智能技术去推动社会进步。
展望未来，随着各行各业数字化转型的步伐加速，数据在未来企业与组织中的作用越来越重要，企业与组织在产品、业务乃至商业模式都需要基于数据而进行重构，数据智能可以说是未来企业与组织所必备的。
“未来十年，将会是数据智能的黄金十年。尤其是数据智能技术与行业业务场景融合，会带来巨大的市场机会。未来市场竞争最重要的是，拥有领先的技术，并且将技术在高价值、高频的场景进行深度融合。相信这个市场一定会诞生出世界级的巨头公司。”苏萌最后表示道。
写在最后
Chairman and CEO Su Meng points is a triathlon enthusiast, he loves a challenge fraught with difficulties. Because of this, he will be "long-distance gene" among the company has injected percentage points during the decade, the face of all the "temptation", points no choice shortcuts, but carries beginning of the heart, down to earth step by step forward. I believe in the future of intelligent market data, points will be able to run out of their own a vast world.