Ramakanth Kavuluru

RASP: efficient multidimensional range query on attack-resilient encrypted databases

Range query is one of the most frequently used queries for online data analytics. Providing such a query service could be expensive for the data owner. With the development of services computing and cloud computing, it has become possible to outsource large databases to database service providers and let the providers maintain the range-query service. With outsourced services, the data owner can greatly reduce the cost in maintaining computing infrastructure and data-rich applications. However, the service provider, although honestly processing queries, may be curious about the hosted data and received queries. Most existing encryption based approaches require linear scan over the entire database, which is inappropriate for online data analytics on large databases. While a few encryption solutions are more focused on efficiency side, they are vulnerable to attackers equipped with certain prior knowledge. We propose the Random Space Encryption (RASP) approach that allows efficient range search with stronger attack resilience than existing efficiency-focused approaches. We use RASP to generate indexable auxiliary data that is resilient to prior knowledge enhanced attacks. Range queries are securely transformed to the encrypted data space and then efficiently processed with a two-stage processing algorithm. We thoroughly studied the potential attacks on the encrypted data and queries at three different levels of prior knowledge available to an attacker. Experimental results on synthetic and real datasets show that this encryption approach allows efficient processing of range queries with high resilience to attacks.

Context-driven automatic subgraph creation for literature-based discovery

Background Literature-based discovery (LBD) is characterized by uncovering hidden associations in non-interacting scientific literature. Prior approaches to LBD include use of: 1) domain expertise and structured background knowledge to manually filter and explore the literature, 2) distributional statistics and graph-theoretic measures to rank interesting connections, and 3) heuristics to help eliminate spurious connections. However, manual approaches to LBD are not scalable and purely distributional approaches may not be sufficient to obtain insights into the meaning of poorly understood associations. While several graph-based approaches have the potential to elucidate associations, their effectiveness has not been fully demonstrated. A considerable degree of a priori knowledge, heuristics, and manual filtering is still required. Objectives In this paper we implement and evaluate a context-driven, automatic subgraph creation method that captures multifaceted complex associations between biomedical concepts to facilitate LBD. Given a pair of concepts, our method automatically generates a ranked list of subgraphs, which provide informative and potentially unknown associations between such concepts. Methods To generate subgraphs, the set of all MEDLINE articles that contain either of the two specified concepts (A, C) are first collected. Then binary relationships or assertions, which are automatically extracted from the MEDLINE articles, called semantic predications, are used to create a labeled directed predications graph. In this predications graph, a path is represented as a sequence of semantic predications. The hierarchical agglomerative clustering (HAC) algorithm is then applied to cluster paths that are bounded by the two concepts (A, C). HAC relies on implicit semantics captured through Medical Subject Heading (MeSH) descriptors, and explicit semantics from the MeSH hierarchy, for clustering. Paths that exceed a threshold of semantic relatedness are clustered into subgraphs based on their shared context. Finally, the automatically generated clusters are provided as a ranked list of subgraphs. Results The subgraphs generated using this approach facilitated the rediscovery of 8 out of 9 existing scientific discoveries. In particular, they directly (or indirectly) led to the recovery of several intermediates (or B-concepts) between A- and C-terms, while also providing insights into the meaning of the associations. Such meaning is derived from predicates between the concepts, as well as the provenance of the semantic predications in MEDLINE. Additionally, by generating subgraphs on different thematic dimensions (such as Cellular Activity, Pharmaceutical Treatment and Tissue Function), the approach may enable a broader understanding of the nature of complex associations between concepts. Finally, in a statistical evaluation to determine the interestingness of the subgraphs, it was observed that an arbitrary association is mentioned in only approximately 4 articles in MEDLINE on average. Conclusion These results suggest that leveraging the implicit and explicit semantics provided by manually assigned MeSH descriptors is an effective representation for capturing the underlying context of complex associations, along multiple thematic dimensions in LBD situations.

An empirical evaluation of supervised learning approaches in assigning diagnosis codes to electronic medical records

BACKGROUND Diagnosis codes are assigned to medical records in healthcare facilities by trained coders by reviewing all physician authored documents associated with a patients visit. This is a necessary and complex task involving coders adhering to coding guidelines and coding all assignable codes. With the popularity of electronic medical records (EMRs), computational approaches to code assignment have been proposed in the recent years. However, most efforts have focused on single and often short clinical narratives, while realistic scenarios warrant full EMR level analysis for code assignment. OBJECTIVE We evaluate supervised learning approaches to automatically assign international classification of diseases (ninth revision) - clinical modification (ICD-9-CM) codes to EMRs by experimenting with a large realistic EMR dataset. The overall goal is to identify methods that offer superior performance in this task when considering such datasets. METHODS We use a dataset of 71,463 EMRs corresponding to in-patient visits with discharge date falling in a two year period (2011-2012) from the University of Kentucky (UKY) Medical Center. We curate a smaller subset of this dataset and also use a third gold standard dataset of radiology reports. We conduct experiments using different problem transformation approaches with feature and data selection components and employing suitable label calibration and ranking methods with novel features involving code co-occurrence frequencies and latent code associations. RESULTS Over all codes with at least 50 training examples we obtain a micro F-score of 0.48. On the set of codes that occur at least in 1% of the two year dataset, we achieve a micro F-score of 0.54. For the smaller radiology report dataset, the classifier chaining approach yields best results. For the smaller subset of the UKY dataset, feature selection, data selection, and label calibration offer best performance. CONCLUSIONS We show that datasets at different scale (size of the EMRs, number of distinct codes) and with different characteristics warrant different learning approaches. For shorter narratives pertaining to a particular medical subdomain (e.g., radiology, pathology), classifier chaining is ideal given the codes are highly related with each other. For realistic in-patient full EMRs, feature and data selection methods offer high performance for smaller datasets. However, for large EMR datasets, we observe that the binary relevance approach with learning-to-rank based code reranking offers the best performance. Regardless of the training dataset size, for general EMRs, label calibration to select the optimal number of labels is an indispensable final step.

Convolutional neural networks for biomedical text classification: application in indexing biomedical articles

Building high accuracy text classifiers is an important task in biomedicine given the wealth of information hidden in unstructured narratives such as research articles and clinical documents. Due to large feature spaces, traditionally, discriminative approaches such as logistic regression and support vector machines with n-gram and semantic features (e.g., named entities) have been used for text classification where additional performance gains are typically made through feature selection and ensemble approaches. In this paper, we demonstrate that a more direct approach using convolutional neural networks (CNNs) outperforms several traditional approaches in biomedical text classification with the specific use-case of assigning medical subject headings (or MeSH terms) to biomedical articles. Trained annotators at the national library of medicine (NLM) assign on an average 13 codes to each biomedical article, thus semantically indexing scientific literature to support NLMs PubMed search system. Recent evidence suggests that effective automated efforts for MeSH term assignment start with binary classifiers for each term. In this paper, we use CNNs to build binary text classifiers and achieve an absolute improvement of over 3% in macro F-score over a set of selected hard-to-classify MeSH terms when compared with the best prior results on a public dataset. Additional experiments on 50 high frequency terms in the dataset also show improvements with CNNs. Our results indicate the strong potential of CNNs in biomedical text classification tasks.

An up-to-date knowledge-based literature search and exploration framework for focused bioscience domains

In domain-specific search systems, knowledge of a domain of interest is embedded as a backbone that guides the search process. But the knowledge used in most such systems 1. exists only for few well known broad domains; 2. is of a basic nature: either purely hierarchical or involves only few relationship types; and 3. is not always kept up-to-date missing insights from recently published results. In this paper we present a framework and implementation of a focused and up-to-date knowledge-based search system, called Scooner, that utilizes domain-specific knowledge extracted from recent bioscience abstracts. To our knowledge, this is the first attempt in the field to address all three shortcomings mentioned above. Since recent introduction for operational use at Applied Biotechnology Branch of AFRL, some biologists are using Scooner on a regular basis, while it is being made available for use by many more. Initial evaluations point to the promise of the approach in addressing the challenge we set out to address.

Leveraging Output Term Co-occurrence Frequencies and Latent Associations in Predicting Medical Subject Headings

Trained indexers at the National Library of Medicine (NLM) manually tag each biomedical abstract with the most suitable terms from the Medical Subject Headings (MeSH) terminology to be indexed by their PubMed information system. MeSH has over 26,000 terms and indexers look at each articles full text while assigning the terms. Recent automated attempts focused on using the article title and abstract text to identify MeSH terms for the corresponding article. Most of these approaches used supervised machine learning techniques that use already indexed articles and the corresponding MeSH terms. In this paper, we present a new indexing approach that leverages term co-occurrence frequencies and latent term associations computed using MeSH term sets corresponding to a set of nearly 18 million articles already indexed with MeSH terms by indexers at NLM. The main goal of our study is to gauge the potential of output label co-occurrences, latent associations, and relationships extracted from free text in both unsupervised and supervised indexing approaches. In this paper, using a novel and purely unsupervised approach, we achieve a micro-F-score that is comparable to those obtained using supervised machine learning techniques. By incorporating term co-occurrence and latent association features into a supervised learning framework, we also improve over the best results published on two public datasets.

Supervised Extraction of Diagnosis Codes from EMRs: Role of Feature Selection, Data Selection, and Probabilistic Thresholding

Extracting diagnosis codes from medical records is a complex task carried out by trained coders by reading all the documents associated with a patients visit. With the popularity of electronic medical records (EMRs), computational approaches to code extraction have been proposed in the recent years. Machine learning approaches to multi-label text classification provide an important methodology in this task given each EMR can be associated with multiple codes. In this paper, we study the the role of feature selection, training data selection, and probabilistic threshold optimization in improving different multi-label classification approaches. We conduct experiments based on two different datasets: a recent gold standard dataset used for this task and a second larger and more complex EMR dataset we curated from the University of Kentucky Medical Center. While conventional approaches achieve results comparable to the state-of-the-art on the gold standard dataset, on our complex in-house dataset, we show that feature selection, training data selection, and probabilistic thresholding provide significant gains in performance.

Unsupervised Medical Subject Heading Assignment Using Output Label Co-occurrence Statistics and Semantic Predications

Librarians at the National Library of Medicine tag each biomedical abstract to be indexed by their Pubmed information system with terms from the Medical Subject Headings (MeSH) terminology. The MeSH terminology has over 26,000 terms and indexers look at each articles full text to assign a set of most suitable terms for indexing it. Several recent automated attempts focused on using the article title and abstract text to identify MeSH terms for the corresponding article. Most of these approaches used supervised machine learning techniques that use already indexed articles and the corresponding MeSH terms. In this paper, we present a novel unsupervised approach using named entity recognition, relationship extraction, and output label co-occurrence frequencies of MeSH term pairs from the existing set of 22 million articles already indexed with MeSH terms by librarians at NLM. The main goal of our study is to gauge the potential of output label co-occurrence statistics and relationships extracted from free text in unsupervised indexing approaches. Especially, in biomedical domains, output label co-occurrences are generally easier to obtain than training data involving document and label set pairs owing to the sensitive nature of textual documents containing protected health information. Our methods achieve a micro F-score that is comparable to those obtained using supervised machine learning techniques with training data consisting of document label set pairs. Baseline comparisons reveal strong prospects for further research in exploiting label co-occurrences and relationships extracted from free text in recommending terms for indexing biomedical articles.

Unsupervised Extraction of Diagnosis Codes from EMRs Using Knowledge-Based and Extractive Text Summarization Techniques

Diagnosis codes are extracted from medical records for billing and reimbursement and for secondary uses such as quality control and cohort identification. In the US, these codes come from the standard terminology ICD-9-CM derived from the international classification of diseases (ICD). ICD-9 codes are generally extracted by trained human coders by reading all artifacts available in a patients medical record following specific coding guidelines. To assist coders in this manual process, this paper proposes an unsupervised ensemble approach to automatically extract ICD-9 diagnosis codes from textual narratives included in electronic medical records (EMRs). Earlier attempts on automatic extraction focused on individual documents such as radiology reports and discharge summaries. Here we use a more realistic dataset and extract ICD-9 codes from EMRs of 1000 inpatient visits at the University of Kentucky Medical Center. Using named entity recognition (NER), graph-based concept-mapping of medical concepts, and extractive text summarization techniques, we achieve an example based average recall of 0.42 with average precision 0.47; compared with a baseline of using only NER, we notice a 12% improvement in recall with the graph-based approach and a 7% improvement in precision using the extractive text summarization approach. Although diagnosis codes are complex concepts often expressed in text with significant long range non-local dependencies, our present work shows the potential of unsupervised methods in extracting a portion of codes. As such, our findings are especially relevant for code extraction tasks where obtaining large amounts of training data is difficult.

2n-Periodic Binary Sequences with Fixed k-Error Linear Complexity for k = 2 or 3

The linear complexity of sequences is an important measure to gauge the cryptographic strength of key streams used in stream ciphers. The instability of linear complexity caused by changing a few symbols of sequences can be measured using k-error linear complexity. In their SETA 2006 paper, Fu, Niederreiter, and Su [3] studied linear complexity and 1-error linear complexity of 2n-periodic binary sequences to characterize such sequences with fixed 1-error linear complexity. In this paper we study the linear complexity and the k-error linear complexity of 2n-periodic binary sequences in a more general setting using a combination of algebraic, combinatorial, and algorithmic methods. This approach allows us to characterize 2n-periodic binary sequences with fixed 2-error or 3-error linear complexity L, when the Hamming weight of the binary representation of 2ni¾? Lis