Active learning: A step towards automating medical concept extraction

Objective This paper presents an automatic active learning-based system for the extraction of medical concepts from clinical free-text reports. Specifically, (1) the contribution of active learning in reducing the annotation effort, and (2) the robustness of incremental active learning framework across different selection criteria and datasets is determined. Materials and methods The comparative performance of an active learning framework and a fully supervised approach were investigated to study how active learning reduces the annotation effort while achieving the same effectiveness as a supervised approach. Conditional Random Fields as the supervised method, and least confidence and information density as two selection criteria for active learning framework were used. The effect of incremental learning vs. standard learning on the robustness of the models within the active learning framework with different selection criteria was also investigated. Two clinical datasets were used for evaluation: the i2b2/VA 2010 NLP challenge and the ShARe/CLEF 2013 eHealth Evaluation Lab. Results The annotation effort saved by active learning to achieve the same effectiveness as supervised learning is up to 77%, 57%, and 46% of the total number of sequences, tokens, and concepts, respectively. Compared to the Random sampling baseline, the saving is at least doubled. Discussion Incremental active learning guarantees robustness across all selection criteria and datasets. The reduction of annotation effort is always above random sampling and longest sequence baselines. Conclusion Incremental active learning is a promising approach for building effective and robust medical concept extraction models, while significantly reducing the burden of manual annotation.

Medical free-text to concept mapping as an information retrieval problem

Concept mapping involves determining relevant concepts from a free-text input, where concepts are defined in an external reference ontology. This is an important process that underpins many applications for clinical information reporting, derivation of phenotypic descriptions, and a number of state-of-the-art medical information retrieval methods. Concept mapping can be cast into an information retrieval (IR) problem: free-text mentions are treated as queries and concepts from a reference ontology as the documents to be indexed and retrieved. This paper presents an empirical investigation applying general-purpose IR techniques for concept mapping in the medical domain. A dataset used for evaluating medical information extraction is adapted to measure the effectiveness of the considered IR approaches. Standard IR approaches used here are contrasted with the effectiveness of two established benchmark methods specifically developed for medical concept mapping. The empirical findings show that the IR approaches are comparable with one benchmark method but well below the best benchmark.

Information extraction from web services : a comparison of Tokenisation algorithms

Most web service discovery systems use keyword-based search algorithms and, although partially successful, sometimes fail to satisfy some users information needs. This has given rise to several semantics-based approaches that look to go beyond simple attribute matching and try to capture the semantics of services. However, the results reported in the literature vary and in many cases are worse than the results obtained by keyword-based systems. We believe the accuracy of the mechanisms used to extract tokens from the non-natural language sections of WSDL files directly affects the performance of these techniques, because some of them can be more sensitive to noise. In this paper three existing tokenization algorithms are evaluated and a new algorithm that outperforms all the algorithms found in the literature is introduced.

Evaluating medical information retrieval

This paper presents a framework for evaluating information retrieval of medical records. We use the BLULab corpus, a large collection of real-world de-identified medical records. The collection has been hand coded by clinical terminol- ogists using the ICD-9 medical classification system. The ICD codes are used to devise queries and relevance judge- ments for this collection. Results of initial test runs using a baseline IR system are provided. Queries and relevance judgements are online to aid further research in medical IR. Please visit: http://koopman.id.au/med_eval.

Graph-based concept weighting for medical information retrieval

This paper presents a graph-based method to weight medical concepts in documents for the purposes of information retrieval. Medical concepts are extracted from free-text documents using a state-of-the-art technique that maps n-grams to concepts from the SNOMED CT medical ontology. In our graph-based concept representation, concepts are vertices in a graph built from a document, edges represent associations between concepts. This representation naturally captures dependencies between concepts, an important requirement for interpreting medical text, and a feature lacking in bag-of-words representations. We apply existing graph-based term weighting methods to weight medical concepts. Using concepts rather than terms addresses vocabulary mismatch as well as encapsulates terms belonging to a single medical entity into a single concept. In addition, we further extend previous graph-based approaches by injecting domain knowledge that estimates the importance of a concept within the global medical domain. Retrieval experiments on the TREC Medical Records collection show our method outperforms both term and concept baselines. More generally, this work provides a means of integrating background knowledge contained in medical ontologies into data-driven information retrieval approaches.

Ontology-based feature modeling for construction information extraction from a building information model

A building information model (BIM) provides a rich representation of a building's design. However, there are many challenges in getting construction-specific information from a BIM, limiting the usability of BIM for construction and other downstream processes. This paper describes a novel approach that utilizes ontology-based feature modeling, automatic feature extraction based on ifcXML, and query processing to extract information relevant to construction practitioners from a given BIM. The feature ontology generically represents construction-specific information that is useful for a broad range of construction management functions. The software prototype uses the ontology to transform the designer-focused BIM into a construction-specific feature-based model (FBM). The formal query methods operate on the FBM to further help construction users to quickly extract the necessary information from a BIM. Our tests demonstrate that this approach provides a richer representation of construction-specific information compared to existing BIM tools.

Multiscale features and information extraction of online strain for long-span bridges

The strain data acquired from structural health monitoring (SHM) systems play an important role in the state monitoring and damage identification of bridges. Due to the environmental complexity of civil structures, a better understanding of the actual strain data will help filling the gap between theoretical/laboratorial results and practical application. In the study, the multi-scale features of strain response are first revealed after abundant investigations on the actual data from two typical long-span bridges. Results show that, strain types at the three typical temporal scales of 10^5, 10^2 and 10^0 sec are caused by temperature change, trains and heavy trucks, and have their respective cut-off frequency in the order of 10^-2, 10^-1 and 10^0 Hz. Multi-resolution analysis and wavelet shrinkage are applied for separating and extracting these strain types. During the above process, two methods for determining thresholds are introduced. The excellent ability of wavelet transform on simultaneously time-frequency analysis leads to an effective information extraction. After extraction, the strain data will be compressed at an attractive ratio. This research may contribute to a further understanding of actual strain data of long-span bridges; also, the proposed extracting methodology is applicable on actual SHM systems.

External knowledge and query strategies in active learning: A study in clinical information extraction

This paper presents a new active learning query strategy for information extraction, called Domain Knowledge Informativeness (DKI). Active learning is often used to reduce the amount of annotation effort required to obtain training data for machine learning algorithms. A key component of an active learning approach is the query strategy, which is used to iteratively select samples for annotation. Knowledge resources have been used in information extraction as a means to derive additional features for sample representation. DKI is, however, the first query strategy that exploits such resources to inform sample selection. To evaluate the merits of DKI, in particular with respect to the reduction in annotation effort that the new query strategy allows to achieve, we conduct a comprehensive empirical comparison of active learning query strategies for information extraction within the clinical domain. The clinical domain was chosen for this work because of the availability of extensive structured knowledge resources which have often been exploited for feature generation. In addition, the clinical domain offers a compelling use case for active learning because of the necessary high costs and hurdles associated with obtaining annotations in this domain. Our experimental findings demonstrated that 1) amongst existing query strategies, the ones based on the classification model’s confidence are a better choice for clinical data as they perform equally well with a much lighter computational load, and 2) significant reductions in annotation effort are achievable by exploiting knowledge resources within active learning query strategies, with up to 14% less tokens and concepts to manually annotate than with state-of-the-art query strategies.

Analysis of word embeddings and sequence features for clinical information extraction

This study investigates the use of unsupervised features derived from word embedding approaches and novel sequence representation approaches for improving clinical information extraction systems. Our results corroborate previous findings that indicate that the use of word embeddings significantly improve the effectiveness of concept extraction models; however, we further determine the influence that the corpora used to generate such features have. We also demonstrate the promise of sequence-based unsupervised features for further improving concept extraction.

Overview of the ShARe/CLEF eHealth Evaluation Lab 2014

This paper reports on the 2nd ShARe/CLEFeHealth evaluation lab which continues our evaluation resource building activities for the medical domain. In this lab we focus on patients' information needs as opposed to the more common campaign focus of the specialised information needs of physicians and other healthcare workers. The usage scenario of the lab is to ease patients and next-of-kins' ease in understanding eHealth information, in particular clinical reports. The 1st ShARe/CLEFeHealth evaluation lab was held in 2013. This lab consisted of three tasks. Task 1 focused on named entity recognition and normalization of disorders; Task 2 on normalization of acronyms/abbreviations; and Task 3 on information retrieval to address questions patients may have when reading clinical reports. This year's lab introduces a new challenge in Task 1 on visual-interactive search and exploration of eHealth data. Its aim is to help patients (or their next-of-kin) in readability issues related to their hospital discharge documents and related information search on the Internet. Task 2 then continues the information extraction work of the 2013 lab, specifically focusing on disorder attribute identification and normalization from clinical text. Finally, this year's Task 3 further extends the 2013 information retrieval task, by cleaning the 2013 document collection and introducing a new query generation method and multilingual queries. De-identified clinical reports used by the three tasks were from US intensive care and originated from the MIMIC II database. Other text documents for Tasks 1 and 3 were from the Internet and originated from the Khresmoi project. Task 2 annotations originated from the ShARe annotations. For Tasks 1 and 3, new annotations, queries, and relevance assessments were created. 50, 79, and 91 people registered their interest in Tasks 1, 2, and 3, respectively. 24 unique teams participated with 1, 10, and 14 teams in Tasks 1, 2 and 3, respectively. The teams were from Africa, Asia, Canada, Europe, and North America. The Task 1 submission, reviewed by 5 expert peers, related to the task evaluation category of Effective use of interaction and targeted the needs of both expert and novice users. The best system had an Accuracy of 0.868 in Task 2a, an F1-score of 0.576 in Task 2b, and Precision at 10 (P@10) of 0.756 in Task 3. The results demonstrate the substantial community interest and capabilities of these systems in making clinical reports easier to understand for patients. The organisers have made data and tools available for future research and development.

Towards semantic search and inference in electronic medical records : an approach using concept-based information retrieval

Background This paper presents a novel approach to searching electronic medical records that is based on concept matching rather than keyword matching. Aim The concept-based approach is intended to overcome specific challenges we identified in searching medical records. Method Queries and documents were transformed from their term-based originals into medical concepts as defined by the SNOMED-CT ontology. Results Evaluation on a real-world collection of medical records showed our concept-based approach outperformed a keyword baseline by 25% in Mean Average Precision. Conclusion The concept-based approach provides a framework for further development of inference based search systems for dealing with medical data.

Exploiting medical hierarchies for concept-based information retrieval

Search technologies are critical to enable clinical sta to rapidly and e ectively access patient information contained in free-text medical records. Medical search is challenging as terms in the query are often general but those in rel- evant documents are very speci c, leading to granularity mismatch. In this paper we propose to tackle granularity mismatch by exploiting subsumption relationships de ned in formal medical domain knowledge resources. In symbolic reasoning, a subsumption (or `is-a') relationship is a parent-child rela- tionship where one concept is a subset of another concept. Subsumed concepts are included in the retrieval function. In addition, we investigate a number of initial methods for combining weights of query concepts and those of subsumed concepts. Subsumption relationships were found to provide strong indication of relevant information; their inclusion in retrieval functions yields performance improvements. This result motivates the development of formal models of rela- tionships between medical concepts for retrieval purposes.

Semantic judgement of medical concepts : combining syntagmatic and paradigmatic information with the tensor encoding model

This paper outlines a novel approach for modelling semantic relationships within medical documents. Medical terminologies contain a rich source of semantic information critical to a number of techniques in medical informatics, including medical information retrieval. Recent research suggests that corpus-driven approaches are effective at automatically capturing semantic similarities between medical concepts, thus making them an attractive option for accessing semantic information. Most previous corpus-driven methods only considered syntagmatic associations. In this paper, we adapt a recent approach that explicitly models both syntagmatic and paradigmatic associations. We show that the implicit similarity between certain medical concepts can only be modelled using paradigmatic associations. In addition, the inclusion of both types of associations overcomes the sensitivity to the training corpus experienced by previous approaches, making our method both more effective and more robust. This finding may have implications for researchers in the area of medical information retrieval.

An evaluation of corpus-driven measures of medical concept similarity for information retrieval

Measures of semantic similarity between medical concepts are central to a number of techniques in medical informatics, including query expansion in medical information retrieval. Previous work has mainly considered thesaurus-based path measures of semantic similarity and has not compared different corpus-driven approaches in depth. We evaluate the effectiveness of eight common corpus-driven measures in capturing semantic relatedness and compare these against human judged concept pairs assessed by medical professionals. Our results show that certain corpus-driven measures correlate strongly (approx 0.8) with human judgements. An important finding is that performance was significantly affected by the choice of corpus used in priming the measure, i.e., used as evidence from which corpus-driven similarities are drawn. This paper provides guidelines for the implementation of semantic similarity measures for medical informatics and concludes with implications for medical information retrieval.

Exploiting inference from semantic annotations for information retrieval

The increasing amount of information that is annotated against standardised semantic resources offers opportunities to incorporate sophisticated levels of reasoning, or inference, into the retrieval process. In this position paper, we reflect on the need to incorporate semantic inference into retrieval (in particular for medical information retrieval) as well as previous attempts that have been made so far with mixed success. Medical information retrieval is a fertile ground for testing inference mechanisms to augment retrieval. The medical domain offers a plethora of carefully curated, structured, semantic resources, along with well established entity extraction and linking tools, and search topics that intuitively require a number of different inferential processes (e.g., conceptual similarity, conceptual implication, etc.). We argue that integrating semantic inference in information retrieval has the potential to uncover a large amount of information that otherwise would be inaccessible; but inference is also risky and, if not used cautiously, can harm retrieval.