Neu-IR: The SIGIR 2016 Workshop on Neural Information Retrieval
Neu-IR: The SIGIR 2016 Workshop on Neural Information Retrieval
复制标题
Neu-IR：SIGIR 2016 神经信息检索研讨会
DOI：
10.1145/2911451.2917762
复制
发表时间：
2016
期刊：
影响因子：
--
通讯作者：
M. D. Rijke
中科院分区：
文献类型：
--
作者：
Nick Craswell;W. B. Croft;Jiafeng Guo;Bhaskar Mitra;M. D. Rijke
近年来，深度神经网络在语音识别和计算机视觉任务方面取得了显着的性能改进，并在自动语音翻译、图像字幕和会话代理等新应用领域取得了令人兴奋的突破。尽管在自然语言处理（NLP）任务（例如，语言建模和机器翻译，深度神经网络在信息检索（IR）任务上的表现相对较少。最近在这一领域的工作主要集中在词嵌入和短文本相似性的神经模型。在信息检索的这一领域缺乏许多积极的成果，部分原因是因为IR任务（如排名）与NLP任务有根本的不同，但也因为IR和神经网络社区才开始关注这些技术在核心信息检索问题上的应用。鉴于深度学习已经产生了如此大的影响，首先是语音处理和计算机视觉，现在越来越多地对计算语言学产生影响，似乎很明显，深度学习将对信息检索产生重大影响，这是一个理想的时间在这个领域举办研讨会。Neu-IR（发音为“新IR”）将是一个与深度学习和其他基于神经网络的IR方法相关的新研究论坛。其目的是为人们提供一个展示新工作和早期成果的机会，比较神经网络工具包的注释，分享最佳实践，并讨论这一研究领域面临的主要挑战。
In recent years, deep neural networks have yielded significant performance improvements on speech recognition and computer vision tasks, as well as led to exciting breakthroughs in novel application areas such as automatic voice translation, image captioning, and conversational agents. Despite demonstrating good performance on natural language processing (NLP) tasks (e.g., language modelling and machine translation, the performance of deep neural networks on information retrieval (IR) tasks has had relatively less scrutiny. Recent work in this area has mainly focused on word embeddings and neural models for short text similarity. The lack of many positive results in this area of information retrieval is partially due to the fact that IR tasks such as ranking are fundamentally different from NLP tasks, but also because the IR and neural network communities are only beginning to focus on the application of these techniques to core information retrieval problems. Given that deep learning has made such a big impact, first on speech processing and computer vision and now, increasingly, also on computational linguistics, it seems clear that deep learning will have a major impact on information retrieval and that this is an ideal time for a workshop in this area. Neu-IR (pronounced "new IR") will be a forum for new research relating to deep learning and other neural network based approaches to IR. The purpose is to provide an opportunity for people to present new work and early results, compare notes on neural network toolkits, share best practices, and discuss the main challenges facing this line of research.
DOI：
--
发表时间：
2015-06
期刊：
ArXiv
影响因子：
--
作者：
O. Vinyals;Quoc V. Le
通讯作者：
O. Vinyals;Quoc V. Le