178 research outputs found
Recent advances in energy efficient query processing
No full textWeb search companies distribute their infrastructures and operations across several, geographically distant data centers. This distributed architecture facilitates high performance query processing, which is fundamental for the success of a Web search engine. At the same time, data centers require an huge amount of electricity to operate their computing resources. In this extended abstract, we briey discuss our recent works for improving the energy efficiency of query processing systems. Firstly, we introduce a novel query forwarding algorithm which exploits green energy sources to reduce the electricity expenditure and carbon footprint of Web search engines. Then, we propose to delegate the CPU power management from a server' operative system directly to the query processing application, to reduce the energy consumption of a search engine's servers. Finally, we introduce PESOS, a scheduling algorithm which manages the CPU power consumption on a per-query basis while considering query latency constraints
Query Embedding Pruning for Dense Retrieval
No full textRecent advances in dense retrieval techniques have offered the promise of being able not just to re-rank documents using contextualised language models such as BERT, but also to use such models to identify documents from the collection in the first place. However, when using dense retrieval approaches that use multiple embedded representations for each query, a large number of documents can be retrieved for each query, hindering the efficiency of the method. Hence, this work is the first to consider efficiency improvements in the context of a dense retrieval approach (namely ColBERT), by pruning query term embeddings that are estimated not to be useful for retrieving relevant documents. Our proposed query embeddings pruning reduces the cost of the dense retrieval operation, as well as reducing the number of documents that are retrieved and hence require to be fully scored. Experiments conducted on the MSMARCO passage ranking corpus demonstrate that, when reducing the number of query embeddings used from 32 to 3 based on the collection frequency of the corresponding tokens, query embedding pruning results in no statistically significant differences in effectiveness, while reducing the number of documents retrieved by 70%. In terms of mean response time for the end-to-end to end system, this results in a 2.65x speedup
A study on query energy consumption in web search engines
No full textCommercial web search engines are usually deployed on data centers, which leverage thousands of servers to eficiently answer queries on a large scale. Thanks to these distributed infrastructures, search engines can quickly serve high query volumes. However, the energy consumed by these many servers poses economical and environmental challenges for the Web search engine companies. To tackle such challenges, we advocate the importance of quantifying the energy consumption of a search engine. Therefore, in this study we experimentally analyze energy consumption on a per query basis. Our aim is to evaluate how much energy is consumed by a search server to answer a single query, i.e, its query energy consumption. To perform such measurements, experiments are conducted using the TREC ClueWeb09 collection and the MSN 2006 query log. Results suggest that solving queries require an amount of energy directly proportional to the query processing time
On Approximate Nearest Neighbour Selection for Multi-Stage Dense Retrieval
No full textDense retrieval, which describes the use of contextualised language models such as BERT to identify documents from a collection by leveraging approximate nearest neighbour (ANN) techniques, has been increasing in popularity. Two families of approaches have emerged, depending on whether documents and queries are represented by single or multiple embeddings. ColBERT, the exemplar of the latter, uses an ANN index and approximate scores to identify a set of candidate documents for each query embedding, which are then re-ranked using accurate document representations. In this manner, a large number of documents can be retrieved for each query, hindering the efficiency of the approach. In this work, we investigate the use of ANN scores for ranking the candidate documents, in order to decrease the number of candidate documents being fully scored. Experiments conducted on the MSMARCO passage ranking corpus demonstrate that, by cutting of the candidate set by using the approximate scores to only 200 documents, we can still obtain an effective ranking without statistically significant differences in effectiveness, and resulting in a 2x speedup in efficiency
Declarative Experimentation in Information Retrieval Using PyTerrier
No full textThe advent of deep machine learning platforms such as Tensorflow and Pytorch, developed in expressive high-level languages such as Python, have allowed more expressive representations of deep neural network architectures. We argue that such a powerful formalism is missing in information retrieval (IR), and propose a framework called PyTerrier that allows advanced retrieval pipelines to be expressed, and evaluated, in a declarative manner close to their conceptual design. Like the aforementioned frameworks that compile deep learning experiments into primitive GPU operations, our framework targets IR platforms as backends in order to execute and evaluate retrieval pipelines. Further, we can automatically optimise the retrieval pipelines to increase their efficiency to suite a particular IR platform backend. Our experiments, conducted on TREC Robust and ClueWeb09 test collections, demonstrate the efficiency benefits of these optimisations for retrieval pipelines involving both the Anserini and Terrier IR platforms
Multiple query processing via logic function factoring
No full textSome extensions to search systems require support for multiple query processing. This is the case with query variations, i.e., differ- ent query formulations of the same information need. The results of their processing can be fused together to improve effectiveness, but this requires to traverse more than once the query terms’ post- ing lists, thus prolonging the multiple query processing time. In this work, we propose an approach to optimize the processing of query variations to reduce their overall response time. Similarly to the standard Boolean model, we firstly represent a group of query variations as a logic function where Boolean variables represent query terms. We then apply factoring to such function, in order to produce a more compact but logically equivalent representation. The factored form is used to process the query variations in a single pass over the inverted index. We experimentally show that our approach can improve by up to 1.95× the mean processing time of a multiple query with no statistically significant degradation in terms of NDCG@10
Energy-Efficient Query Processing in Web Search Engines
No full textWeb search engines are composed by thousands of query processing nodes, i.e., servers dedicated to process user queries. Such many servers consume a significant amount of energy, mostly accountable to their CPUs, but they are necessary to ensure low latencies, since users expect sub-second response times (e.g., 500 ms). However, users can hardly notice response times that are faster than their expectations. Hence, we propose the Predictive Energy Saving Online Scheduling Algorithm (PESOS) to select the most appropriate CPU frequency to process a query on a per-core basis. PESOS aims at process queries by their deadlines, and leverage high-level scheduling information to reduce the CPU energy consumption of a query processing node. PESOS bases its decision on query efficiency predictors, estimating the processing volume and processing time of a query. We experimentally evaluate PESOS upon the TREC ClueWeb09B collection and the MSN2006 query log. Results show that PESOS can reduce the CPU energy consumption of a query processing node up to 20 percent energy saving, while the competitor requires a fine parameter tuning and it may incurs in uncontrollable latency violations
Upper Bound Approximations for BlockMaxWand
No full textBlockMaxWand is a recent advance on the Wand dynamic pruning
technique, which allows efficient retrieval without any effectiveness
degradation to rank K. However, while BMW uses docid-sorted indices,
it relies on recording the upper bound of the term weighting
model scores for each block of postings in the inverted index. Such
a requirement can be disadvantageous in situations such as when
an index must be updated. In this work, we examine the appropriateness
of upper-bound approximation – which have previously
been shown suitable for Wand– in providing efficient retrieval for
BMW. Experiments on the ClueWeb12 category B13 corpus using
5000 queries from a real search engine’s query log demonstrate that
BMW still provides benefits w.r.t. Wand when approximate upper
bounds are used, and that, if approximations on upper bounds are
tight, BMW with approximate upper bounds can provide efficiency
gains w.r.t.Wand with exact upper bounds, in particular for queries
of short to medium length
- …
