Tabular data provide answers to a significant portion of search queries. However, reciting an entire result table is impractical in conversational search systems. We propose to generate natural language summaries as answers to describe the complex information contained in a table. Through crowdsourcing experiments, we build a new conversation-oriented, open-domain table summarization dataset. It includes annotated table summaries, which not only answer questions but also help people explore other information in the table. We utilize this dataset to develop automatic table summarization systems as SOTA baselines. Based on the experimental results, we identify challenges and point out future research directions that this resource will support.

Term frequency is a common method for identifying the importance of a term in a document. But term frequency ignores how a term interacts with its text context, which is key to estimating document-specific term weights. This paper proposes a Deep Contextualized Term Weighting framework (DeepCT) that maps the contextualized term representations from BERT to into context-aware term weights for passage retrieval. The new, deep term weights can be stored in an ordinary inverted index for efficient retrieval. Experiments on two datasets demonstrate that DeepCT greatly improves the accuracy of first-stage passage retrieval algorithms.

Language model pre-training has spurred a great deal of attention for tasks involving natural language understanding, and has been successfully applied to many downstream tasks with impressive results. Within information retrieval, many of these solutions are too costly to stand on their own, requiring multi-stage ranking architectures. Recent work has begun to consider how to "backport" salient aspects of these computationally expensive models to previous stages of the retrieval pipeline. One such instance is DeepCT, which uses BERT to re-weight term importance in a given context at the passage level. This process, which is computed offline, results in an augmented inverted index with re-weighted term frequency values. In this work, we conduct an investigation of query processing efficiency over DeepCT indexes. Using a number of candidate generation algorithms, we reveal how term re-weighting can impact query processing latency, and explore how DeepCT can be used as a static index pruning technique to accelerate query processing without harming search effectiveness.

Bag-of-words document representations play a fundamental role in modern search engines, but their power is limited by the shallow frequency-based term weighting scheme. This paper proposes HDCT, a context-aware document term weighting framework for document indexing and retrieval. It first estimates the semantic importance of a term in the context of each passage. These fine-grained term weights are then aggregated into a document-level bag-of-words representation, which can be stored into a standard inverted index for efficient retrieval. This paper also proposes two approaches that enable training HDCT without relevance labels. Experiments show that an index using HDCT weights significantly improved the retrieval accuracy compared to typical term-frequency and state-of-the-art embedding-based indexes.

Neural networks provide new possibilities to automatically learn complex language patterns and query-document relations. Neural IR models have achieved promising results in learning query-document relevance patterns, but few explorations have been done on understanding the text content of a query or a document. This paper studies leveraging a recently-proposed contextual neural language model, BERT, to provide deeper text understanding for IR. Experimental results demonstrate that the contextual text representations from BERT are more effective than traditional word embeddings. Compared to bag-of-words retrieval models, the contextual language model can better leverage language structures, bringing large improvements on queries written in natural languages. Combining the text understanding ability with search knowledge leads to an enhanced pre-trained BERT model that can benefit related search tasks where training data are limited.

Finding diagrams that contain a specific part or a similar part is important in many engineering tasks. In this search task, the query part is expected to match only a small region in a complex image. This paper investigates several local matching networks that explicitly model local region-to-region similarities. Deep convolutional neural networks extract local features and model local matching patterns. Spatial convolution is employed to cross-match local regions at different scale levels, addressing cases where the target part appears at a different scale, position, and/or angle. A gating network automatically learns region importance, removing noise from sparse areas and visual metadata in engineering diagrams.

This paper studies the consistency of the kernel-based neural ranking model K-NRM, a recent state-of-the-art neural IR model, which is important for reproducible research and deployment in the industry. We find that K-NRM has low variance on relevance-based metrics across experimental trials. In spite of this low variance in overall performance, different trials produce different document rankings for individual queries. The main source of variance in our experiments was found to be different latent matching patterns captured by K-NRM. In the IR-customized word embeddings learned by K-NRM, the query-document word pairs follow two different matching patterns that are equally effective, but align word pairs differently in the embedding space. The different latent matching patterns enable a simple yet effective approach to construct ensemble rankers, which improve K-NRM's effectiveness and generalization abilities.

This paper presents \textttConv-KNRM, a Convolutional Kernel-based Neural Ranking Model that models n-gram soft matches for ad-hoc search. Instead of exact matching query and document n-grams, \textttConv-KNRM uses Convolutional Neural Networks to represent n-grams of various lengths and soft matches them in a unified embedding space. The n-gram soft matches are then utilized by the kernel pooling and learning-to-rank layers to generate the final ranking score. \textttConv-KNRM can be learned end-to-end and fully optimized from user feedback. The learned model»s generalizability is investigated by testing how well it performs in a related domain with small amounts of training data. Experiments on English search logs, Chinese search logs, and TREC Web track tasks demonstrated consistent advantages of \textttConv-KNRM over prior neural IR methods and feature-based methods.

We present a learning-to-rank approach for resource selection. We develop features for resource ranking and present a training approach that does not require human judgments. Our method is well-suited to environments with a large number of resources such as selective search, is an improvement over the state-of-the-art in resource selection for selective search, and is statistically equivalent to exhaustive search even for recall-oriented metrics such as [email protected], an area in which selective search was lacking.

This paper proposes K-NRM, a kernel based neural model for document ranking. Given a query and a set of documents, K-NRM uses a translation matrix that models word-level similarities via word embeddings, a new kernel-pooling technique that uses kernels to extract multi-level soft match features, and a learning-to-rank layer that combines those features into the final ranking score. The whole model is trained end-to-end. The ranking layer learns desired feature patterns from the pairwise ranking loss. The kernels transfer the feature patterns into soft-match targets at each similarity level and enforce them on the translation matrix. The word embeddings are tuned accordingly so that they can produce the desired soft matches. Experiments on a commercial search engine's query log demonstrate the improvements of K-NRM over prior feature-based and neural-based states-of-the-art, and explain the source of K-NRM's advantage: Its kernel-guided embedding encodes a similarity metric tailored for matching query words to document words, and provides effective multi-level soft matches.

Selective search is a cluster-based distributed retrieval architecture that reduces computational costs by partitioning a corpus into topical shards, and selectively searching them. Prior research formed topical shards by clustering the corpus based on the documents' contents. This content-based partitioning strategy reveals common topics in a corpus. However, the topic distribution produced by clustering may not match the distribution of topics in search traffic, which may reduce the effectiveness of selective search.

Selective distributed search is a retrieval architecture that reduces search costs by partitioning a corpus into topical shards such that only a few shards need to be searched for each query. Prior research created topical shards by using random seed documents to cluster a random sample of the full corpus. The resource selection algorithm might use a different random sample of the corpus. These random components make selective search non-deterministic. This paper studies how these random components affect experimental results. Experiments on two ClueWeb09 corpora and four query sets show that in spite of random components, selective search is stable for most queries.