微信AI拿下NLP竞赛全球冠军,“二孩”智言团队的实习生立功了

3 月 1 日 量子位

乾明 发自 凹非寺 
量子位 报道 | 公众号 QbitAI

又是中国团队夺冠全球AI竞赛。

这次来自腾讯,源自微信AI,拿下NLP领域全球桂冠。

近日,在第七届对话系统技术挑战赛(DSTC7)上,首次亮相的微信智言团队一路过关斩将,最终拿下冠军。

DSTC7挑战赛,由来自微软研究院、卡耐基梅隆大学的科学家与2013年发起,分为三个赛道。

来自微信模式识别中心的微信智言参加的是其中之一,基于Fact(比如百科文章、Blog评论)与对话上下文信息,自动生成回答。

在这次比赛中,微信智言提出了一种基于多方位注意力机制,能够“阅读” Fact与对话上下文信息,并利用原创动态聚类解码器,产生与Fact和上下文相关并且有趣的回答。

在自动和人工评测都取得最佳成绩,力压赛道中其他竞争对手。

但外界可能有所不知的是,如此成绩背后,向来低调著称的微信,原动力主要为锻炼队伍。

微信智言团队告诉量子位,核心参赛组员,仅一名来自CMU的T4专家为指导,一名实习生为主力,最后为微信证明实力。

或者更具体来说——为微信智言证明实力。

微信AI“二孩”

微信智言,是继微信智聆之后,微信团队推出的又一AI技术品牌,专注于智能对话和自然语言处理等技术的研究与应用,被称为微信AI“二孩”。

在2017年12月开始筹备,2019年1月份微信公开课正式亮相。与腾讯AI Lab相比,微信智言更加聚焦于NLP领域,不论是文本分类、问答、对话,还是语义解析,都有大量投入。

虽然亮相较晚,但实力一点不差。

目前,微信智言团队由来自微软和CMU的T4专家坐镇,共有10多名成员,全部来自国外内顶尖高校的硕士及博士,且都具备学术研究能力与开发能力。

微信智言透露,团队每两个月都会请第三方机构对团队的对话系统实力进行评估。

结果表明,从对话、用户满意度角度来说,微信智言基本上处于第一梯队。

不过,微信智言也强调:竞赛成绩能够展现研究实力,但团队最主要重心,始终在业务上。

瞄准四大领域

微信智言的目标,是打造“对话即服务”平台。现已在智能硬件、PaaS、行业云和AI Bot等四大领域进行了业务部署。

智能硬件

在这一领域,微信智言推出了技术平台“小微”,并面向开发者提供硬件SDK、云服务和APP端接入的一站式开发集成方案。

接入小微平台后,设备可通过智能对话使用音乐、新闻和视频等内置技能,也可通过开放技能平台创建配置自定义技能。

目前已有近百款硬件产品接入小微对话系统,比如哈曼旗下的的JBL等,服务范围涉及运动户外、家居生活、智能车载、智慧产业和新兴市场等领域。

PaaS&行业云

与腾讯云合作,微信智言搭建了PaaS和行业云的语义服务。

其中,PaaS是以云服务为基础的智能语义平台,第三方开发者可以通过API搭建自己的应用和服务,基于行业特点构建出不同的服务和能力。代表性的案例有香格里拉和雅朵合作的智能酒店应用等等。

行业云面向企业客户提供完整的解决方案。为企业快速搭建智能客服平台和行业任务智能对话系统。代表案例有微信支付和春秋航空等。

AI Bot

在微信智言的计划中,未来将提供基于微信生态的语音助手服务。

微信智言表示,AI Bot不仅仅是一个语音助手,还是用户在微信生态下链接内容和服务的入口。不仅能够听懂语音指令,还能将微信生态中的优质内容和服务推送给你。

微信智言成长进行时

虽然已经取得了不小的成果,但其仍旧在成长的过程中。

接下来,微信智言主要发力三个方向:

  • 第一,服务好外部第三方的音箱等智能设备市场,以音箱市场为排头兵,不断迭代,提升效果,满足用户需求。

  • 第二,为腾讯内部业务提供支持,和各个BG合作, 提升产品体验。

  • 第三,继续在学术领域对真正有价值的问题进行探索和解决,产学研结合,将微信智言打造成国内NLP领域的第一梯队。

本次参加竞赛,就是为了在学界进一步锤炼技术水平,同时吸引更多人才加盟。

而且微信也用了“心思”,比如带队实习生的导师,是腾讯技术级别不低的T4专家,诚意之至,可见一斑。

另外,微信智言方面也说了,这只是一个小小体现,更多惊喜,欢迎你亲自去发现。

最后,附上微信智言团队的论文传送门:

Cluster-based Beam Search for Pointer-Generator Chatbot Grounded by Knowledge
http://workshop.colips.org/dstc7/papers/03.pdf



加入社群

量子位现开放「AI+行业」社群,面向AI行业相关从业者,技术、产品等人员,根据所在行业可选择相应行业社群,在量子位公众号(QbitAI)对话界面回复关键词“行业群”,获取入群方式。行业群会有审核,敬请谅解。


此外,量子位AI社群正在招募,欢迎对AI感兴趣的同学,在量子位公众号(QbitAI)对话界面回复关键字“交流群”,获取入群方式。

诚挚招聘

量子位正在招募编辑/记者,工作地点在北京中关村。期待有才气、有热情的同学加入我们!相关细节,请在量子位公众号(QbitAI)对话界面,回复“招聘”两个字。

量子位 QbitAI · 头条号签约作者

վ'ᴗ' ի 追踪AI技术和产品新动态

喜欢就点「好看」吧 !

登录查看更多
点赞 0

This paper takes a step towards theoretical analysis of the relationship between word embeddings and context embeddings in models such as word2vec. We start from basic probabilistic assumptions on the nature of word vectors, context vectors, and text generation. These assumptions are well supported either empirically or theoretically by the existing literature. Next, we show that under these assumptions the widely-used word-word PMI matrix is approximately a random symmetric Gaussian ensemble. This, in turn, implies that context vectors are reflections of word vectors in approximately half the dimensions. As a direct application of our result, we suggest a theoretically grounded way of tying weights in the SGNS model.

点赞 0
阅读1+

Rapidly developed neural models have achieved competitive performance in Chinese word segmentation (CWS) as their traditional counterparts. However, most of methods encounter the computational inefficiency especially for long sentences because of the increasing model complexity and slower decoders. This paper presents a simple neural segmenter which directly labels the gap existence between adjacent characters to alleviate the existing drawback. Our segmenter is fully end-to-end and capable of performing segmentation very fast. We also show a performance difference with different tag sets. The experiments show that our segmenter can provide comparable performance with state-of-the-art.

点赞 0
阅读1+

Often when multiple labels are obtained for a training example it is assumed that there is an element of noise that must be accounted for. It has been shown that this disagreement can be considered signal instead of noise. In this work we investigate using soft labels for training data to improve generalization in machine learning models. However, using soft labels for training Deep Neural Networks (DNNs) is not practical due to the costs involved in obtaining multiple labels for large data sets. We propose soft label memorization-generalization (SLMG), a fine-tuning approach to using soft labels for training DNNs. We assume that differences in labels provided by human annotators represent ambiguity about the true label instead of noise. Experiments with SLMG demonstrate improved generalization performance on the Natural Language Inference (NLI) task. Our experiments show that by injecting a small percentage of soft label training data (0.03% of training set size) we can improve generalization performance over several baselines.

点赞 0
阅读1+

The field of natural language processing has seen impressive progress in recent years, with neural network models replacing many of the traditional systems. A plethora of new models have been proposed, many of which are thought to be opaque compared to their feature-rich counterparts. This has led researchers to analyze, interpret, and evaluate neural networks in novel and more fine-grained ways. In this survey paper, we review analysis methods in neural language processing, categorize them according to prominent research trends, highlight existing limitations, and point to potential directions for future work.

点赞 0
阅读1+
Top