Vasant G. Honavar

Learn++: an incremental learning algorithm for supervised neural networks

We introduce Learn++, an algorithm for incremental training of neural network (NN) pattern classifiers. The proposed algorithm enables supervised NN paradigms, such as the multilayer perceptron (MLP), to accommodate new data, including examples that correspond to previously unseen classes. Furthermore, the algorithm does not require access to previously used data during subsequent incremental learning sessions, yet at the same time, it does not forget previously acquired knowledge. Learn++ utilizes ensemble of classifiers by generating multiple hypotheses using training data sampled according to carefully tailored distributions. The outputs of the resulting classifiers are combined using a weighted majority voting procedure. We present simulation results on several benchmark datasets as well as a real-world classification task. Initial results indicate that the proposed algorithm works rather well in practice. A theoretical upper bound on the error of the classifiers constructed by Learn++ is also provided.

Mobile health technology evaluation: the mHealth evidence workshop.

Creative use of new mobile and wearable health information and sensing technologies (mHealth) has the potential to reduce the cost of health care and improve well-being in numerous ways. These applications are being developed in a variety of domains, but rigorous research is needed to examine the potential, as well as the challenges, of utilizing mobile technologies to improve health outcomes. Currently, evidence is sparse for the efficacy of mHealth. Although these technologies may be appealing and seemingly innocuous, research is needed to assess when, where, and for whom mHealth devices, apps, and systems are efficacious. In order to outline an approach to evidence generation in the field of mHealth that would ensure research is conducted on a rigorous empirical and theoretic foundation, on August 16, 2011, researchers gathered for the mHealth Evidence Workshop at NIH. The current paper presents the results of the workshop. Although the discussions at the meeting were cross-cutting, the areas covered can be categorized broadly into three areas: (1) evaluating assessments; (2) evaluating interventions; and (3) reshaping evidence generation using mHealth. This paper brings these concepts together to describe current evaluation standards, discuss future possibilities, and set a grand goal for the emerging field of mHealth research.

Predicting linear B-cell epitopes using string kernels

The identification and characterization of B‐cell epitopes play an important role in vaccine design, immunodiagnostic tests, and antibody production. Therefore, computational tools for reliably predicting linear B‐cell epitopes are highly desirable. We evaluated Support Vector Machine (SVM) classifiers trained utilizing five different kernel methods using fivefold cross‐validation on a homology‐reduced data set of 701 linear B‐cell epitopes, extracted from Bcipep database, and 701 non‐epitopes, randomly extracted from SwissProt sequences. Based on the results of our computational experiments, we propose BCPred, a novel method for predicting linear B‐cell epitopes using the subsequence kernel. We show that the predictive performance of BCPred (AUC = 0.758) outperforms 11 SVM‐based classifiers developed and evaluated in our experiments as well as our implementation of AAP (AUC = 0.7), a recently proposed method for predicting linear B‐cell epitopes using amino acid pair antigenicity. Furthermore, we compared BCPred with AAP and ABCPred, a method that uses recurrent neural networks, using two data sets of unique B‐cell epitopes that had been previously used to evaluate ABCPred. Analysis of the data sets used and the results of this comparison show that conclusions about the relative performance of different B‐cell epitope prediction methods drawn on the basis of experiments using data sets of unique B‐cell epitopes are likely to yield overly optimistic estimates of performance of evaluated methods. This argues for the use of carefully homology‐reduced data sets in comparing B‐cell epitope prediction methods to avoid misleading conclusions about how different methods compare to each other. Our homology‐reduced data set and implementations of BCPred as well as the APP method are publicly available through our web‐based server, BCPREDS, at: http://ailab.cs.iastate.edu/bcpreds/. Copyright

Constructive neural-network learning algorithms for pattern classification

Constructive learning algorithms offer an attractive approach for the incremental construction of near-minimal neural-network architectures for pattern classification. They help overcome the need for ad hoc and often inappropriate choices of network topology in algorithms that search for suitable weights in a priori fixed network architectures. Several such algorithms are proposed in the literature and shown to converge to zero classification errors (under certain assumptions) on tasks that involve learning a binary to binary mapping (i.e., classification problems involving binary-valued input attributes and two output categories). We present two constructive learning algorithms MPyramid-real and MTiling-real that extend the pyramid and tiling algorithms, respectively, for learning real to M-ary mappings (i.e., classification problems involving real-valued input attributes and multiple output classes). We prove the convergence of these algorithms and empirically demonstrate their applicability to practical pattern classification problems. Additionally, we show how the incorporation of a local pruning step can eliminate several redundant neurons from MTiling-real networks.

Intelligent agents for intrusion detection

The paper focuses on intrusion detection and countermeasures with respect to widely-used operating systems and networks. The design and architecture of an intrusion detection system built from distributed agents is proposed to implement an intelligent system on which data mining can be performed to provide global, temporal views of an entire networked system. A starting point for agent intelligence in the system is the research into the use of machine learning over system call traces from the privileged sendmail program on UNIX. The authors use a rule learning algorithm to classify the system call traces for intrusion detection purposes and show the results.

RNABindR: a server for analyzing and predicting RNA-binding sites in proteins

Understanding interactions between proteins and RNA is key to deciphering the mechanisms of many important biological processes. Here we describe RNABindR, a web-based server that identifies and displays RNA-binding residues in known protein–RNA complexes and predicts RNA-binding residues in proteins of unknown structure. RNABindR uses a distance cutoff to identify which amino acids contact RNA in solved complex structures (from the Protein Data Bank) and provides a labeled amino acid sequence and a Jmol graphical viewer in which RNA-binding residues are displayed in the context of the three-dimensional structure. Alternatively, RNABindR can use a Naive Bayes classifier trained on a non-redundant set of protein–RNA complexes from the PDB to predict which amino acids in a protein sequence of unknown structure are most likely to bind RNA. RNABindR automatically displays ‘high specificity’ and ‘high sensitivity’ predictions of RNA-binding residues. RNABindR is freely available at http://bindr.gdcb.iastate.edu/RNABindR.

Lightweight agents for intrusion detection

We have designed and implemented an intrusion detection system (IDS) prototype based on mobile agents. Our agents travel between monitored systems in a network of distributed systems, obtain information from data cleaning agents, classify and correlate information, and report the information to a user interface and database via mediators.Agent systems with lightweight agent support allow runtime addition of new capabilities to agents. We describe the design of our Multi-agent IDS and show how lightweight agent capabilities allowed us to add communication and collaboration capabilities to the mobile agents in our IDS.

Predicting DNA-binding sites of proteins from amino acid sequence

BackgroundUnderstanding the molecular details of protein-DNA interactions is critical for deciphering the mechanisms of gene regulation. We present a machine learning approach for the identification of amino acid residues involved in protein-DNA interactions.ResultsWe start with a Naïve Bayes classifier trained to predict whether a given amino acid residue is a DNA-binding residue based on its identity and the identities of its sequence neighbors. The input to the classifier consists of the identities of the target residue and 4 sequence neighbors on each side of the target residue. The classifier is trained and evaluated (using leave-one-out cross-validation) on a non-redundant set of 171 proteins. Our results indicate the feasibility of identifying interface residues based on local sequence information. The classifier achieves 71% overall accuracy with a correlation coefficient of 0.24, 35% specificity and 53% sensitivity in identifying interface residues as evaluated by leave-one-out cross-validation. We show that the performance of the classifier is improved by using sequence entropy of the target residue (the entropy of the corresponding column in multiple alignment obtained by aligning the target sequence with its sequence homologs) as additional input. The classifier achieves 78% overall accuracy with a correlation coefficient of 0.28, 44% specificity and 41% sensitivity in identifying interface residues. Examination of the predictions in the context of 3-dimensional structures of proteins demonstrates the effectiveness of this method in identifying DNA-binding sites from sequence information. In 33% (56 out of 171) of the proteins, the classifier identifies the interaction sites by correctly recognizing at least half of the interface residues. In 87% (149 out of 171) of the proteins, the classifier correctly identifies at least 20% of the interface residues. This suggests the possibility of using such classifiers to identify potential DNA-binding motifs and to gain potentially useful insights into sequence correlates of protein-DNA interactions.ConclusionNaïve Bayes classifiers trained to identify DNA-binding residues using sequence information offer a computationally efficient approach to identifying putative DNA-binding sites in DNA-binding proteins and recognizing potential DNA-binding motifs.

A two-stage classifier for identification of protein--protein interface residues

MOTIVATION The ability to identify protein-protein interaction sites and to detect specific amino acid residues that contribute to the specificity and affinity of protein interactions has important implications for problems ranging from rational drug design to analysis of metabolic and signal transduction networks. RESULTS We have developed a two-stage method consisting of a support vector machine (SVM) and a Bayesian classifier for predicting surface residues of a protein that participate in protein-protein interactions. This approach exploits the fact that interface residues tend to form clusters in the primary amino acid sequence. Our results show that the proposed two-stage classifier outperforms previously published sequence-based methods for predicting interface residues. We also present results obtained using the two-stage classifier on an independent test set of seven CAPRI (Critical Assessment of PRedicted Interactions) targets. The success of the predictions is validated by examining the predictions in the context of the three-dimensional structures of protein complexes.

A framework for semantic web services discovery

This paper describes a framework for ontology-based flexible discovery of Semantic Web services. The proposed approach relies on user-supplied, context-specific mappings from an user ontology to relevant domain ontologies used to specify Web services. We show how a users query for a Web service that meets certain selection criteria can be transformed into queries that can be processed by a matchmaking engine that is aware of the relevant domain ontologies and Web services. We also describe how user-specified preferences for Web services in terms of non-functional requirements (e.g., QoS) can be incorporated into the Web service discovery mechanism to generate a partially ordered list of services that meet user-specified functional requirements.