Learning to predict expression efficacy of vectors in recombinant protein production.

A spectral graph theoretic approach to quantification and calibration of collective morphological differences in cell images.

Fusion of systems for automated cell phenotype image classification.



Workshop summary Automated interpretation and modelling of cell images.


Feature Space Transformation for Semi%2DSupervised Learning for Protein Subcellular Localization in Fluorescence Microscopy Images.


Global and componentwise extrapolations for accelerating training of Bayesian networks and conditional random fields.

Periodic step%2Dsize adaptation in second%2Dorder gradient descent for single%2Dpass on%2Dline structured learning.

Integrating high dimensional bi%2Ddirectional parsing models for gene mention tagging.

Integrating high dimensional bi%2Ddirectional parsing models for gene mention tagging.

Cross%2Dlingual audio%2Dto%2Dtext alignment for multimedia content management.

Training Conditional Random Fields by Periodic Step Size Adaptation for Large%2DScale Text Mining.
For applications with consecutive incoming training examples, on-line learning has the potential to achieve a likelihood as high as off-line learning without scanning all available training examples and usually has a much smaller memory footprint. To train CRFs on-line, this paper presents the Periodic Step size Adaptation (PSA) method to dynamically adjust the learning rates in stochastic gradient descent. We applied our method to three large scale text mining tasks. Experimental results show that PSA outperforms the best off-line algorithm, L-BFGS, by many hundred times, and outperforms the best on-line algorithm, SMD, by an order of magnitude in terms of the number of passes required to scan the training data set. 1.
Analysis and Enhancement of Conditional Random Fields Gene Mention Taggers in BioCreative II Challenge Evaluation.
Background: Tagging gene and gene product mentions in scientific text is an important initial step of literature mining. In BioCreative 2 challenge, the conditional random fields model (CRF) was the most prevailing method in the gene mention task. In this paper, we analyze two best performing CRF-based systems in BioCreative 2. We examine their key claims and propose enhancement based on the analysis results. Results: We implemented their systems in MALLET as specified in their report and in CRF++, a different CRF package, to empirically analyze their claims. We found that their feature set is effective for models trained by MALLET, but a smaller set works better for those by CRF++. We confirmed the effectiveness of pairing parentheses as a post processing step. We found that backward parsing is not always superior to forward parsing. The benefit of applying bidirectional parsing is the creation of a wider variety of complementary models. We elaborated the notion of divergent models by relating it to the difference of the increments of ture positives and false positives of the union model. Conclusions: To further enhance the performance, we can integrate more models based on the elaborated notion of divergent models that we derived to minimize the number of models required. Background At present, scientific literature is still the largest and most reliable source of biomedical knowledge. A great
Gene Name Service No%2DNonsense Alias Resolution Service for Homo Sapiens Genes.



Boosting multiclass learning with repeating codes and weak detectors for protein subcellular localization.

Global and Componentwise Extrapolation for Accelerating Data Mining from Large Incomplete Data Sets with the EM Algorithm.
The Expectation-Maximization (EM) algorithm is one of the most popular algorithms for data mining from incomplete data. However, when applied to large data sets with a large proportion of missing data, the EM algorithm may converge slowly. The triple jump extrapolation method can effectively accelerate the EM algorithm by substantially reducing the number of iterations required for EM to converge. There are two options for the triple jump method, global extrapolation (TJEM) and componentwise extrapolation (CTJEM). We tried these two methods for a variety of probabilistic models and found that in general, global extraplolation yields a better performance, but there are cases where componentwise extrapolation yields very high speedup. In this paper, we investigate when componentwise extrapolation should be preferred. We conclude that, when the Jacobian of the EM mapping is diagonal or block diagonal, CTJEM should be preferred. We show how to determine whether a Jacobian is diagonal or block diagonal and experimentally confirm our claim. In particular, we show that CTJEM is especially effective for the semi-supervised Bayesian classifier model given a highly sparse data set. 1.
Language Identification by Using Syllable%2DBased Duration Classification on Code%2DSwitching Speech.

FASTSNP an always up%2Dto%2Ddate and extendable service for SNP function analysis and prioritization.

A Recommender for Targeted Advertisement of Unsought Products in E%2DCommerce.

Triple Jump Acceleration for the EM Algorithm.
This paper presents the triple jump framework for accelerating the EM algorithm and other bound optimization methods. The idea is to extrapolate the third search point based on the previous two search points found by regular EM. As the convergence rate of regular EM becomes slower, the distance of the triple jump will be longer, and thus provide higher speedup for data sets where EM converges slowly. Experimental results show that the triple jump framework significantly outperforms EM and other acceleration methods of EM for a variety of probabilistic models, especially when the data set is sparse or less structured, which usually slow down EM but are common in real world data sets. The results also show that the triple jump framework is particularly effective for Cluster Models. 1.
Reconfigurable Web wrapper agents for biological information integration.

Mining Skewed and Sparse Transaction Data for Personalized Shopping Recommendation.

Reconfigurable Web Wrapper Agents for Web Information Integration.
This paper provides a solution for rapidly building software agents that can serve as Web wrappers for biological information integration. We define an XML-based language called WNDL, which provides a representation of a Web browsing session.
The Hybrid Poisson Aspect Model for Personalized Shopping Recommendation.

Reconfigurable Web Wrapper Agents for Web Information Integration.
This paper provides a solution for rapidly building software agents that can serve as Web wrappers for biological information integration. We define an XML-based language called WNDL, which provides a representation of a Web browsing session.
Query Answering Using Ontologies in Agent%2Dbased Resource Sharing Environment for Biological Web Information Integrating.

Automatic information extraction from semi%2Dstructured Web pages by pattern discovery.
Workshop held in conjunction withPreface Today on-line communities, as well as individuals, produce a substantial amount of unstructured (and extemporaneous) content, arising from tacit and explicit knowledge sharing. Various approaches, both in the managerial and computer science fields, are seeking ways to crystallize the- somewhat volatile, but often highly valuable- knowledge contained in communities &#034;chatters&#034;. Among those approaches, the most relevants appear to be those aimed at developing and formalizing agreed-upon semantic representations of specific knowledge domains (i.e. domain ontologies). Nonetheless, the intrinsic limits of technologies underpinning such approaches tend to push communities members towards the spontaneous adoption of less cumbersome tools, usually offered in the framework of the Web 2.0 (e.g. folksonomies, XML-based tagging, etc.), for sharing and retrieving knowledge. Inside this landscape, community members should be able to access and browse community knowledge transparently and in a personalized way, through tools that
Reconfigurable Web Wrapper Agents.
This paper provides a solution for rapidly building software agents that can serve as Web wrappers for biological information integration. We define an XML-based language called WNDL, which provides a representation of a Web browsing session.
View Inference for Heterogeneous XML Information Integration.

Implications of the Dirichlet Assumption for Discretization of Continuous Variables in Naive Bayesian Classifiers.
Abstract. In a naive Bayesian classifier, discrete variables as well as discretized continuous variables are assumed to have Dirichlet priors. This paper describes the implications and applications of this model selection choice. We start by reviewing key properties of Dirichlet distributions. Among these properties, the most important one is “perfect aggregation, ” which allows us to explain why discretization works for a naive Bayesian classifier. Since perfect aggregation holds for Dirichlets, we can explain that in general, discretization can outperform parameter estimation assuming a normal distribution. In addition, we can explain why a wide variety of well-known discretization methods, such as entropy-based, ten-bin, and binlogl, can perform well with insignificant difference. We designed experiments to verify our explanation using synthesized and real data sets and showed that in addition to well-known methods, a wide variety of discretization methods all perform similarly. Our analysis leads to a lazy discretization method, which discretizes continuous variables according to test data. The Dirichlet assumption implies that lazy methods can perform as well as eager discretization methods. We empirically confirmed this implication and extended the lazy method to classify set-valued and multi-interval data with a naive Bayesian classifier.
Recognizing 100 Speakers Using Homologous Naive Bayes.
This paper presents an extension of the naive Bayesian classi  er, called \homologous naive Bayes (HNB),&#034; which is applied to the  problem of text-independent, close-set speaker recognition. Unlike the  standard naive Bayes, HNB can take advantage of the prior information  that a sequence of input feature vectors belongs to the same unknown  class. We refer to such a sequence a homologous set, which is naturally  available in speaker recognition. We empirically compare HNB with the  Gaussian mixture model (GMM), the most widely used approach to  speaker recognition. Results show that, in spite of its simplisity, HNB can  achieve comparable classi  cation accuracies for up to a hundred speakers  while taking much less resources in terms of time and code size for both  training and classi  cation.
Bayesian classification for data from the same unknown class.

The ANNIGMA%2Dwrapper approach to fast feature selection for neural nets.
This paper presents a novel feature selection approach for backprop neural networks. Previously, an feature selection technique known as the wrapper model was shown effective for decision trees induction. However, it is prohibitively expensive when applied to real-world neural net training characterized by large volumes of data and many feature choices. Our approach incorporates a weight analysis based heuristic called ANNIGMA to direct the search in the wrapper model and allows effective feature selection feasible for neural net applications. Experimental results on standard data sets show that this approach can efficiently reduce the number of features while maintaining or even improving the accuracy. We also report two sucessful applications of our approach in the helicopter maintenance applications.

Ontology Integration in XML.


Semantic Query Optimization for Query Plans of Heterogeneous Multidatabase Systems.
New applications of information systems, such as electronic commerce and healthcare information  systems, need to integrate a large number of heterogeneous databases over computer  networks. Answering a query in these applications usually involves selecting relevant information  sources and generating a query plan to combine the data automatically. As significant progress  has been made in source selection and plan generation, the critical issue has been shifting to  query optimization. This paper presents a semantic query optimization (SQO) approach to  optimizing query plans of heterogeneous multidatabase systems. This approachprovides global  optimization for query plans as well as local optimization for subqueries that retrieve data from  individual database sources. An important feature of our local optimization algorithm is that  weprove necessary and sufficient conditions to eliminate an unnecessary join in a conjunctive  query of arbitrary join topology. This feature allows our...
Discovering Robust Knowledge from Databases that Change.
Many applications of knowledge discovery and data mining such as rule discovery  for semantic query optimization, database integration and decision support, require  the knowledge to be consistent with data. However, databases usually change over  time and makemachine-discovered knowledge inconsistent. Useful knowledge should  be robust against database changessothatitisunlikely to become inconsistentafter  database changes. This paper defines this notion of robustness in the context  of relational databases that contain multiple relations and describes how robustness of  first-order Horn-clause rules can be estimated and applied in knowledge discovery.Our  experiments show that the estimation approach can accurately predict the robustness  of a rule.  
Generating Finite%2DState Transducers for Semi%2DStructured Data Extraction from the Web.
Integrating a large number of Web information sources may significantly increase the utility of the World-Wide Web. A promising solution to the integration is through the use of a Web Information mediator that provides seamless, transparent access for the clients. Information mediators need wrappers to access a Web source as a structured database, but building wrappers by hand is impractical. Previous work on wrapper induction is too restrictive to handle a large number of Web pages that contain tuples with missing attributes, multiple values, variant attribute permutations, exceptions and typos. This paper presents SoftMealy, a novel wrapper representation formalism. This representation is based on a finite-state transducer (FST) and contextual rules. This approach can wrap a wide range of semistructured Web pages because FSTs can encode each different attribute permutation as a path. A SoftMealy wrapper can be induced from a handful of labeled examples using our generalization algori...
Discovering Robust Knowledge from Dynamic Closed World Data.
Many applications of knowledge discovery require  the knowledge to be consistent with data. Examples  include discovering rules for query optimization,  database integration, decision support, etc.  However, databases usually change over time and  make machine-discovered knowledge inconsistent  with data. Useful knowledge should be robust  against database changes so that it is unlikely  to become inconsistent after database changes.  This paper defines this notion of robustness, describes  how to estimate the robustness of Hornclause  rules in closed-world databases, and describes  how the robustness estimation can be applied  in rule discovery systems.  Introduction  Databases are evolving entities. Knowledge discovered from one database state may become invalid or inconsistent with a new database state. Manyapplications  require discovered knowledge to be consistent with the data. Examples are the problem of learning for database query optimization, database integration, knowledge d...
Using Inductive Learning To Generate Rules for Semantic Query Optimization.
Semantic query optimization can dramatically speed up database query answering byknowledge intensive reformulation. But the problem of how to learn the required semantic rules has not been previously solved. This chapter presents a learning approach to solving this problem. In our approach, the learning is triggered by user queries. Then the system uses an inductive learning algorithm to generate semantic rules. This inductive learning algorithm can automatically select useful join paths and attributes to construct rules from a database with many relations. The learned semantic rules are effective for optimization because they will match query patterns and reflect data regularities. Experimental results show that this approach learns sufficient rules for optimization that produces a substantial cost reduction.   
Estimating the Robustness of Discovered Knowledge.
This paper introduces a new measurement, robustness,  to measure the qualityofmachine-discovered knowledge  from real-world databases that change over time.  A piece of knowledge is robust if it is unlikely to  become inconsistent with new database states. Robustness  is different from predictive accuracy in that  by the latter, the system considers only the consistency  of a rule with unseen data, while by the former,  the consistency after deletions and updates of existing  data is also considered. Combining robustness with  other utility measurements, a system can makeintelligent  decisions in learning and maintenance of knowledge  learned from changing databases. This paper  defines robustness, then presents an estimation approach  for the robustness of Horn-clause rules learned  from a relational database. The estimation approach  applies the Laplace law of succession, which can be  efficiently computed. The estimation is based on  database schemas and transaction logs. No domains...
Cooperating Agents for Information Retrieval.
With the vast number of information resources available today, a critical problem is how to locate, retrieve and process information. It would be impractical to build a single unified system that combines all of these information resources. A more promising approach is to build specialized information retrieval agents that provide access to a subset of the information resources and can send requests to other information retrieval agents when appropriate. In this paper we present the architecture of the individual information retrieval agents and describe how this architecture supports a network of cooperating information agents. We describe how these information agents represent their knowledge, communicate with other agents, dynamically construct information retrieval plans, and learn about other agents to improve efficiency. We have already built a small network of agents that have these capabilities and provide access to information for transportation planning.
Rule Introduction for Semantic Query Optimization.

Rule Induction for Semantic Query Optimization.
Semantic query optimization can dramatically speed up database query answering by knowledge intensive reformulation. But the problem of how to learn required semantic rules has not previously been solved. This paper describes an approach using an inductive learning algorithm to solve the problem. In our approach, learning is triggered by user queries and then the system induces semantic rules from the information in databases. The inductive learning algorithm used in this approach can select an appropriate set of relevant attributes from a potentially huge number of attributes in real-world databases. Experimental results demonstrate that this approach can learn sufficient background knowledge to reformulate queries and provide a 57 percent average performance improvement.  1 INTRODUCTION  Speeding up a system&#039;s performance is one of the major goals of machine learning. Explanation-based learning is typically used for speedup learning, while applications of inductive learning are usual...
Reformulating Query Plans for Multidatabase Systems.
A practical heterogeneous, distributed multidatabase system must answer queries efficiently. Conventional query optimization techniques are not adequate here because these techniques are dependent on the database structure, and rely on limited information which is not  sufficient in complicated multidatabase queries. This paper presents an automated approach to reformulating query plans to improve the efficiency of multidatabase queries. This approach uses database abstractions, the knowledge about the contents of databases, to reformulate a query plan into less expensive but semantically equivalent one. We present two algorithms. The first algorithm reformulates subqueries to individual databases, the second algorithm extends the first one and reformulates the entire query plan. Empirical results show that the reformulations can provide significant savings with minimal overhead. The reformulation approach provides a global reduction in the amount of  the intermediate data as well as local opt...
Retrieving and Integrating Data from Multiple Information Sources.
With the current explosion of data, retrieving and integrating information  from various sources is a critical problem. Work in multidatabase systems  has begun to address this problem, but it has primarily focused on methods  for communicating between databases and requires significant effort for each  new database added to the system. This paper describes a more general  approach that exploits a semantic model of a problem domain to integrate  the information from various information sources. The information sources  handled include both databases and knowledge bases, and other information  sources (e.g., programs) could potentially be incorporated into the system.  This paper describes how both the domain and the information sources are  modeled, shows how a query at the domain level is mapped into a set of  queries to individual information sources, and presents algorithms for automatically  improving the efficiency of queries using knowledge about both the  domain and the informat...
