Machine learning: Trends, perspectives, and prospects.

Machine learning addresses the question of how to build computers that improve automatically through experience. It is one of today's most rapidly growing technical fields, lying at the intersection of computer science and statistics, and at the core of artificial intelligence and data science. Recent progress in machine learning has been driven both by the development of new learning algorithms and theory and by the ongoing explosion in the availability of online data and low-cost computation. The adoption of data-intensive machine-learning methods can be found throughout science, technology and commerce, leading to more evidence-based decision-making across many walks of life, including health care, manufacturing, education, financial modeling, policing, and marketing.

Statistical modeling of biomedical corpora: mining the Caenorhabditis Genetic Center Bibliography for genes related to life span.

BACKGROUND: The statistical modeling of biomedical corpora could yield integrated, coarse-to-fine views of biological phenomena that complement discoveries made from analysis of molecular sequence and profiling data. Here, the potential of such modeling is demonstrated by examining the 5,225 free-text items in the Caenorhabditis Genetic Center (CGC) Bibliography using techniques from statistical information retrieval. Items in the CGC biomedical text corpus were modeled using the Latent Dirichlet Allocation (LDA) model. LDA is a hierarchical Bayesian model which represents a document as a random mixture over latent topics; each topic is characterized by a distribution over words. RESULTS: An LDA model estimated from CGC items had better predictive performance than two standard models (unigram and mixture of unigrams) trained using the same data. To illustrate the practical utility of LDA models of biomedical corpora, a trained CGC LDA model was used for a retrospective study of nematode genes known to be associated with life span modification. Corpus-, document-, and word-level LDA parameters were combined with terms from the Gene Ontology to enhance the explanatory value of the CGC LDA model, and to suggest additional candidates for age-related genes. A novel, pairwise document similarity measure based on the posterior distribution on the topic simplex was formulated and used to search the CGC database for "homologs" of a "query" document discussing the life span-modifying clk-2 gene. Inspection of these document homologs enabled and facilitated the production of hypotheses about the function and role of clk-2. CONCLUSION: Like other graphical models for genetic, genomic and other types of biological data, LDA provides a method for extracting unanticipated insights and generating predictions amenable to subsequent experimental validation.

Kernel-based data fusion and its application to protein function prediction in yeast.

Kernel methods provide a principled framework in which to represent many types of data, including vectors, strings, trees and graphs. As such, these methods are useful for drawing inferences about biological phenomena. We describe a method for combining multiple kernel representations in an optimal fashion, by formulating the problem as a convex optimization problem that can be solved using semidefinite programming techniques. The method is applied to the problem of predicting yeast protein functional classifications using a support vector machine (SVM) trained on five types of data. For this problem, the new method performs better than a previously-described Markov random field method, and better than the SVM trained on any single type of data.

Robust sparse hyperplane classifiers: application to uncertain molecular profiling data.

Molecular profiling studies can generate abundance measurements for thousands of transcripts, proteins, metabolites, or other species in, for example, normal and tumor tissue samples. Treating such measurements as features and the samples as labeled data points, sparse hyperplanes provide a statistical methodology for classifying data points into one of two categories (classification and prediction) and defining a small subset of discriminatory features (relevant feature identification). However, this and other extant classification methods address only implicitly the issue of observed data being a combination of underlying signals and noise. Recently, robust optimization has emerged as a powerful framework for handling uncertain data explicitly. Here, ideas from this field are exploited to develop robust sparse hyperplanes, i.e., classification and relevant feature identification algorithms that are resilient to variation in the data. Specifically, each data point is associated with an explicit data uncertainty model in the form of an ellipsoid parameterized by a center and covariance matrix. The task of learning a robust sparse hyperplane from such data is formulated as a second order cone program (SOCP). Gaussian and distribution-free data uncertainty models are shown to yield SOCPs that are equivalent to the SCOP based on ellipsoidal uncertainty. The real-world utility of robust sparse hyperplanes is demonstrated via retrospective analysis of breast cancer related transcript profiles. Data-dependent heuristics are used to compute the parameters of each ellipsoidal data uncertainty model. The generalization performance of a specific implementation, designated "robust LIKNON," is better than its nominal counterpart. Finally, the strengths and limitations of robust sparse hyperplanes are discussed.

Integrated analysis of transcript profiling and protein sequence data.

Transcript profiling can be used to elucidate the molecular and cellular mechanisms involved in ageing and cancer. A recent study of human gastrointestinal stromal tumours (GISTs) with mutations in the KIT gene, Cancer Res. 61 (2001) 8624 exemplifies a common type of investigation. cDNA microarrays were used to generate measurements for 1987 clones in two types of tissues: 13 KIT mutation-positive GISTs and 6 spindle cell tumours from locations outside the gastrointestinal tract. Statistical problems associated with such two-class, high-dimensional profiling data include simultaneous classification and relevant feature identification, probabilistic clustering and protein sequence family modelling. Here, the GIST data were reexamined using specific solutions to these problems, namely sparse hyperplanes, nai;ve Bayes models and profile hidden Markov models respectively. The integrated analysis of molecular profiling and sequence data highlighted 6 clones that may be of clinical and experimental interest. The protein encoded by one of these putative biomarkers defined a novel protein family present in diverse eucarya. The family may be involved in chromosome segregation and/or stability. One family member is a potential biomarker identified recently from a retrospective analysis of transcript profiles for sporadic breast cancer samples from patients with poor and good prognosis, Signal Process. (in press).

Attractor dynamics in feedforward neural networks.

We study the probabilistic generative models parameterized by feedforward neural networks. An attractor dynamics for probabilistic inference in these models is derived from a mean field approximation for large, layered sigmoidal networks. Fixed points of the dynamics correspond to solutions of the mean field equations, which relate the statistics of each unit to those of its Markov blanket. We establish global convergence of the dynamics by providing a Lyapunov function and show that the dynamics generate the signals required for unsupervised learning. Our results for feedforward networks provide a counterpart to those of Cohen-Grossberg and Hopfield for symmetric networks.

Smoothness maximization along a predefined path accurately predicts the speed profiles of complex arm movements.

The speed profiles of arm movements display a number of regularities, including bell-shaped speed profiles in straight reaching movements and an inverse relationship between speed and curvature in extemporaneous drawing movements (described as a 2/3 power law). Here we propose a new model that simultaneously accounts for both regularities by replacing the 2/3 power law with a smoothness constraint. For a given path of the hand in space, our model assumes that the speed profile will be the one that minimizes the third derivative of position (or "jerk"). Analysis of the mathematical relationship between this smoothness constraint and the 2/3 power law revealed that in both two and three dimensions, the power law is equivalent to setting the jerk along the normal to the path to zero; it generates speed predictions that are similar, but clearly distinguishable from the predictions of our model. We have assessed the accuracy of the model on a number of motor tasks in two and three dimensions, involving discrete movements along arbitrary paths, traced with different limb segments. The new model provides a very close fit to the observed speed profiles in all cases. Its performance is uniformly better compared with all existing versions of the 2/3 power law, suggesting that the correlation between speed and curvature may be a consequence of an underlying motor strategy to produce smooth movements. Our results indicate that the relationship between the path and the speed profile of a complex arm movement is stronger than previously thought, especially within a single trial. The accuracy of the model was quite uniform over movements of different shape, size, and average speed. We did not find evidence for segmentation, yet prediction error increased with movement duration, suggesting a continuous fluctuation of the "tempo" of discrete movements. The implications of these findings for motor planning and on-line control are discussed.

The role of inertial sensitivity in motor planning.

To achieve a given motor task a single trajectory must be chosen from the infinite set of possibilities consistent with the task. To investigate such motor planning in a natural environment, we examined the kinematics of reaching movements made around a visual obstacle in three-dimensional space. Within each session, the start and end points of the movement were uniformly varied around the obstacle. However, the distribution of the near points, where the paths came closest to the obstacle, showed a strong anisotropy, clustering at the poles of a preferred axis through the center of the obstacle. The preferred axes for movements made with the left and right arms were mirror symmetric about the midsagittal plane, suggesting that the anisotropy stems from intrinsic properties of the arm rather than extrinsic visual factors. One account of these results is a sensitivity model of motor planning, in which the movement path is skewed so that when the hand passes closest to the obstacle, the arm is in a configuration that is least sensitive to perturbations that might cause collision. To test this idea, we measured the mobility ellipse of the arm. The mobility minor axis represents the direction in which the hand is most inertially stable to a force perturbation. In agreement with the sensitivity model, the mobility minor axis was not significantly different from the preferred near point axis. The results suggest that the sensitivity of the arm to perturbations, as determined by its inertial stability, is taken into account in the planning process.

Sensorimotor adaptation in speech production.

Human subjects are known to adapt their motor behavior to a shift of the visual field brought about by wearing prism glasses over their eyes. The analog of this phenomenon was studied in the speech domain. By use of a device that can feed back transformed speech signals in real time, subjects were exposed to phonetically sensible, online perturbations of their own speech patterns. It was found that speakers learn to adjust their production of a vowel to compensate for feedback alterations that change the vowel's perceived phonetic identity; moreover, the effect generalizes across phonetic contexts and to different vowels.

Obstacle avoidance and a perturbation sensitivity model for motor planning.

A novel obstacle avoidance paradigm was used to investigate the planning of human reaching movements. We explored whether the CNS plans arm movements based entirely on the visual space kinematics of the movements, or whether the planning process incorporates specific details of the biomechanical plant to optimize the trajectory plan. Participants reached around an obstacle, the tip of which remained fixed in space throughout the experiment. When the obstacle and the start and target locations were rotated about the tip of the obstacle, the visually specified task constraints retained a rotational symmetry. If movements are planned in visual space, as indicated from a variety of studies on planar point-to-point movements, the resulting trajectories should also be rotationally symmetric across trials. However, systematic variations in movement path were observed as the orientation of the obstacle was changed. These path asymmetries can be accounted for by a class of models in which the planner reduces the likelihood of collision with the obstacle by taking into account the anisotropic sensitivity of the arm to external perturbations or uncertainty in joint level control or proprioception. The model that best matches the experimental results uses planning criteria based on the inertial properties of the arm.

Probabilistic independence networks for hidden Markov probability models.

Graphical techniques for modeling the dependencies of random variables have been explored in a variety of different areas, including statistics, statistical physics, artificial intelligence, speech recognition, image processing, and genetics. Formalisms for manipulating these models have been developed relatively independently in these research communities. In this paper we explore hidden Markov models (HMMs) and related structures within the general framework of probabilistic independence networks (PINs). The paper presents a self-contained review of the basic principles of PINs. It is shown that the well-known forward-backward (F-B) and Viterbi algorithms for HMMs are special cases of more general inference algorithms for arbitrary PINs. Furthermore, the existence of inference and estimation algorithms for more general graphical models provides a set of analysis tools for HMM practitioners who wish to explore a richer class of HMM structures. Examples of relatively complex models to handle sensor fusion and coarticulation in speech recognition are introduced and treated within the graphical model framework to illustrate the advantages of the general approach.

Generalization to local remappings of the visuomotor coordinate transformation.

During visually guided movement, visual representations of target location must be transformed into coordinates appropriate for movement. To investigate the representation and plasticity of the visuomotor coordinate transformation, we examined the changes in pointing behavior after local visuomotor remappings. The visual feedback of finger position was limited to one or two locations in the workspace, at which a discrepancy was introduced between the actual and visually perceived finger position. These remappings induced changes in pointing, which were largest near the locus of remapping and decreased away from it. This pattern of spatial generalization highly constrains models of the computation of the visuomotor transformation in the CNS. A simple model, in which the transformation is computed via the population activity of a set of units with large sensory receptive fields, is shown to capture the observed pattern.

Local linear perceptrons for classification.

A structure composed of local linear perceptrons for approximating global class discriminants is investigated. Such local linear models may be combined in a cooperative or competitive way. In the cooperative model, a weighted sum of the outputs of the local perceptrons is computed where the weight is a function of the distance between the input and the position of the local perceptron. In the competitive model, the cost function dictates a mixture model where only one of the local perceptrons give output. Learning of the local models' positions and the linear mappings they implement are coupled and both supervised. We show that this is preferable to the uncoupled case where the positions are trained in an unsupervised manner before the separate, supervised training of mappings. We use goodness criteria based on the cross-entropy and give learning equations for both the cooperative and competitive cases. The coupled and uncoupled versions of cooperative and competitive approaches are compared among themselves and with multilayer perceptrons of sigmoidal hidden units and radial basis functions (RBFs) of Gaussian units on the application of recognition of handwritten digits. The criteria of comparison are the generalization accuracy, learning time, and the number of free parameters. We conclude that even on such a high-dimensional problem, such local models are promising. They generalize much better than RBF's and use much less memory. When compared with multilayer perceptrons, we note that local models learn much faster and generalize as well and sometimes better with comparable number of parameters.

An internal model for sensorimotor integration.

On the basis of computational studies it has been proposed that the central nervous system internally simulates the dynamic behavior of the motor system in planning, control, and learning; the existence and use of such an internal model is still under debate. A sensorimotor integration task was investigated in which participants estimated the location of one of their hands at the end of movements made in the dark and under externally imposed forces. The temporal propagation of errors in this task was analyzed within the theoretical framework of optimal state estimation. These results provide direct support for the existence of an internal model.

The Organization of Action Sequences: Evidence From a Relearning Task.

In two studies, the organization of sequential behavior in transcription typing was investigated. The design of the studies made it possible to test the hypothesis that sequential skill in typing resides only at an abstract, effector-independent level. Skilled typists (N = 12) learned to type on an altered keyboard in an experimental paradigm that allowed only certain components of the motor control system to adapt to the alterations. When performance was compared on a pretest and a posttest, various decrements in the typists' speed and accuracy were observed. The forms of these decrements provided evidence against a strong form of the effector-independent hypothesis.

Are arm trajectories planned in kinematic or dynamic coordinates? An adaptation study.

There are several invariant features of point-to-point human arm movements: trajectories tend to be straight, smooth, and have bell-shaped velocity profiles. One approach to accounting for these data is via optimization theory; a movement is specified implicitly as the optimum of a cost function, e.g., integrated jerk or torque change. Optimization models of trajectory planning, as well as models not phrased in the optimization framework, generally fall into two main groups-those specified in kinematic coordinates and those specified in dynamic coordinates. To distinguish between these two possibilities we have studied the effects of artificial visual feedback on planar two-joint arm movements. During self-paced point-to-point arm movements the visual feedback of hand position was altered so as to increase the perceived curvature of the movement. The perturbation was zero at both ends of the movement and reached a maximum at the midpoint of the movement. Cost functions specified by hand coordinate kinematics predict adaptation to increased curvature so as to reduce the visual curvature, while dynamically specified cost functions predict no adaptation in the underlying trajectory planner, provided the final goal of the movement can still be achieved. We also studied the effects of reducing the perceived curvature in transverse movements, which are normally slightly curved. Adaptation should be seen in this condition only if the desired trajectory is both specified in kinematic coordinates and actually curved. Increasing the perceived curvature of normally straight sagittal movements led to significant (P < 0.001) corrective adaptation in the curvature of the actual hand movement; the hand movement became curved, thereby reducing the visually perceived curvature. Increasing the curvature of the normally curved transverse movements produced a significant (P < 0.01) corrective adaptation; the hand movement became straighter, thereby again reducing the visually perceived curvature. When the curvature of naturally curved transverse movements was reduced, there was no significant adaptation (P > 0.05). The results of the curvature-increasing study suggest that trajectories are planned in visually based kinematic coordinates. The results of the curvature-reducing study suggest that the desired trajectory is straight in visual space. These results are incompatible with purely dynamic-based models such as the minimum torque change model. We suggest that spatial perception--as mediated by vision--plays a fundamental role in trajectory planning.

Perceptual distortion contributes to the curvature of human reaching movements.

Unconstrained point-to-point human arm movements are generally gently curved, a fact which has been used to assess the validity of models of trajectory formation. In this study we examined the relationship between curvature perception and movement curvature for planar sagittal and transverse arm movements. We found a significant correlation (P < 0.0001, n = 16) between the curvature perceived as straight and the curvature of actual arm movements. We suggest that subjects try to make straight-line movements, but that actual movements are curved because visual perceptual distortion makes the movements appear to be straighter than they really are. We conclude that perceptual distortion of curvature contributes to the curvature seen in human point-to-point arm movements and that this must be taken into account in the assessment of models of trajectory formation.

A model of the learning of arm trajectories from spatial deviations.

Abstract Unconstrained point-to-point reaching movements performed in the horizontal plane tend to follow roughly straight hand paths with smooth, bell-shaped velocity profiles. The objective of the research reported here was to explore the hypothesis that these data reflect an underlying learning process that prefers simple paths in space. Under this hypothesis, movements are learned based only on spatial errors between the actual hand path and a desired hand path; temporally varying targets are not allowed. We designed a neural network architecture that learned to produce neural commands to a set of muscle-like actuators based only on information about spatial errors. Following repetitive executions of the reaching task, the network was able to generate point-to-point horizontal arm movements and the resulting muscle activation patterns and hand trajectories were found to be similar to those observed experimentally for human subjects. The implications of our results with respect to current theories of multijoint limb movement generation are discussed.

The Cascade Neural Network Model and a Speed-Accuracy Trade-Off of Arm Movement.

We propose a hybrid neural network model of aimed arm movements that consists of a feedforward controller and a postural controller. The cascade neural network of Kawato, Maeda, Uno, and Suzuki (1990) was employed as a computational implementation of the feedforward controller. This network computes feedforward motor commands based on a minimum torque-change criterion. If the weighting parameter of the smoothness criterion is fixed and the number of relaxation iterations is rather small, the cascade model cannot calculate the exact torque, and the hand does not reach the desired target by using the feedforward control alone. Thus, one observes an error between the final position and the desired target location. By using a fixed weighting parameter value and a limited iteration number to simulate target-directed arm movements, we found that the cascade model generated a planning time-accuracy trade-off, and a quasi-power-law type of speed-accuracy trade-off. The model provides a candidate neural mechanism to explain the stochastic variability of the time course of the feedforward motor command. Our approach also accounts for several invariant features of multijoint arm trajectories, such as roughly straight hand paths and bell-shaped speed profiles.

Trading relations between tongue-body raising and lip rounding in production of the vowel /u/: a pilot "motor equivalence" study.

Articulatory and acoustic data were used to explore the following hypothesis for the vowel /u/: The objective of articulatory movements is an acoustic goal; varying and reciprocal contributions of different articulators may help to constrain acoustic variation in achieving the goal. Previous articulatory studies of similar hypotheses, expressed entirely in articulatory terms, have been confounded by interdependencies of the variables being studied (e.g., lip and mandible displacements). One case in which this problem may be minimized is that of lip rounding and tongue-body raising (formation of a velo-palatal constriction) for the vowel /u/. Lip rounding and tongue-body raising should have similar acoustic effects for /u/, mainly to lower F2. In multiple repetitions, reciprocal contributions of lip rounding and tongue-body raising could help limit F2 variability for /u/; thus this experiment looked for complementary covariation (negative correlations) in measures of these two parameters. An electro-magnetic midsagittal articulometer (EMMA) was used to track movements of midsagittal points on the tongue body, upper and lower lips, and mandible for large numbers of repetitions of utterances containing /u/. (Interpretation of the data was aided by results from area-function-to-formant modeling.) Three of four subjects showed weak negative correlations, tentatively supporting the hypothesis; a fourth showed the opposite pattern: positive correlations of lip rounding and tongue raising. The results are discussed with respect to ideas about motor equivalence, the nature of speech motor programming, and potential improvements to the paradigm.