Is my model too complex? Evaluating model formulation using model reduction

While mechanistic models tend to be detailed, they are less detailed than the real systems they seek to describe, so judgements are being made about the appropriate level of detail within the process of model development. These judgements are difficult to test, consequently it is easy for models to become over-parameterised, potentially increasing uncertainty in predictions. The work we describe is a step towards addressing these difficulties. We propose and implement a method which explores a family of simpler models obtained by replacing model variables with constants (model reduction by variable replacement). The procedure iteratively searches the simpler model formulations and compares models in terms of their ability to predict observed data, evaluated within a Bayesian framework. The results can be summarised as posterior model probabilities and replacement probabilities for individual variables which lend themselves to mechanistic interpretation. This provides powerful diagnostic information to support model development, and can identify areas of model over-parameterisation with implications for interpretation of model results. We present the application of the method to 3 example models. In each case reduced models are identified which outperform the original full model in terms of comparisons to observations, suggesting some over-parameterisation has occurred during model development. We argue that the proposed approach is relevant to anyone involved in the development or use of process based mathematical models, especially those where understanding is encoded via empirically based relationships.     

Active spline model: A shape based model—interactive segmentation

We propose to formulate point distribution model in terms of centripetal-parameterized Catmull–Rom spline, so that the model-based segmentation is augmented to permit quick edit, and the consequent shape is independent of scale. We train the model in a fashion similar to active shape model, but with fewer salient/landmark points. We use gradient vector flow field as the external force field to drive the segmentation, but we did not adopt the procedures panned out by Cootes et al. to update a shape. Instead, we transform the shape back and forth between model scale and image scale to get the shape converged to the object of interest. To test the method, we turned the solution into an automated algorithm to segment lung on chest radiographs, and achieved an average overlap of 0.879. With edit, the average overlap increased to 0.945, with a minimum of 0.925. The method can be applied on a variety of images, as illustrated in Appendix C. The source code of the algorithm and the demo video can be located at http://jenh.co/2014/01/09/active-spline-models/.     

The subspace Gaussian mixture model—A structured model for speech recognition

We describe a new approach to speech recognition, in which all Hidden Markov Model (HMM) states share the same Gaussian Mixture Model (GMM) structure with the same number of Gaussians in each state. The model is defined by vectors associated with each state with a dimension of, say, 50, together with a global mapping from this vector space to the space of parameters of the GMM. This model appears to give better results than a conventional model, and the extra structure offers many new opportunities for modeling innovations while maintaining compatibility with most standard techniques.     

Study on joint Bayesian model selection and parameter estimation method of GTD model

The Bayesian method is applied to the joint model selection and parameter estimation problem of the GTD model. An algorithm based on RJ-MCMC is designed. This algorithm not only improves the model order selection and parameter estimation accuracy by exploiting the priori information of the GTD model, but also solves the mixed parameter estimation problem of the GTD model properly. Its performance is tested using numerical simulations and data generated by electromagnetic code. It is shown that it gives good model order selection and parameter estimation results, especially for low SNR, closely-spaced components and short data situations.     

Agents’ model of uncertainty

Multi-agent systems play an increasing role in sensor networks, software engineering, web design, e-commerce, robotics, and many others areas. Uncertainty is a fundamental property of these areas. Agent-based systems use probabilistic and other uncertainty models developed earlier without explicit consideration of agents. This paper explores the impact of agents on uncertainty models and theories. We compare two methods of introducing agents to uncertainty theories and propose a new theory called the agent-based uncertainty theory (AUT). We show advantages of AUT for advancing multi-agent systems and for solving an internal fundamental question of uncertainty theories, that is identifying coherent approaches to uncertainty. The advantages of AUT are that it provides a uniform agent-based representation and an operational empirical interpretation for several uncertainty theories such as rough set theory, fuzzy sets theory, evidence theory, and probability theory. We show also that the introduction of agents to intuitionist uncertainty formalisms can reduce their conceptual complexity. To build such uniformity the AUT exploits the fact that agents as independent entities can give conflicting evaluations of the same attribute. The AUT is based on complex aggregations of crisp (non-fuzzy) conflicting judgments of agents. The generality of AUT is derived from the logical classification of types (orders) of conflicts in the agent populations. At the first order of conflict, the two agent populations are disjoint and there is no interference of logic values assigned to any statement p and its negation by agents. The second order of conflict models superposition (interference) of logic values for overlapping agent populations where an agent assigns conflicting logic values (true, false) to the same attribute simultaneously.

Dynamic output feedback model predictive control for nonlinear systems represented by Hammerstein–Wiener model

This paper presents dynamic output feedback model predictive control (DOFMPC) for nonlinear systems represented by a Hammerstein–Wiener model. Compared with a previous work (IET-OFMPC: output feedback model predictive control for nonlinear systems represented by Hammerstein–Wiener model. IET Control Theory & Applications, 2007, 1 (5) pp. 1302–1310), this paper uses the notion of quadratic boundedness to specify the closed-loop stability and guarantees the recursive feasibility of the optimization problems. By optimizing all the parameters of the dynamic output feedback law within a single optimization problem, the computational burden is very huge. Hence, an alternative formulation is also proposed with much lower on-line computational burden. Numerical examples are given to illustrate the effectiveness of the controllers.     

Universally composable anonymous Hash certification model

Ideal function is the fundamental component in the universally composable security model. However, the certification ideal function defined in the universally composable security model realizes the identity authentication by binding identity to messages and the signature, which fails to characterize the special security requirements of anonymous authentication with other kind of certificate. Therefore, inspired by the work of Marten, an anonymous hash certification ideal function and a more universal certificate CA model are proposed in this paper. We define the security requirements and security notions for this model in the framework of universal composable security and prove in the plain model (not in the random-oracle model) that these security notions can be achieved using combinations of a secure digital signature scheme, a symmetrical encryption mechanism, a family of pseudorandom functions, and a family of one-way collision-free hash functions. Considering the limitation of wireless environment and computation ability of wireless devices, this anonymous Hash certification ideal function is realized by using symmetry primitives.     

Does model consistency in business model innovation matter? A contingency‐based approach

Successful business model innovation requires managers to come up not only with new, but also with viable business models. To this end, it has been argued that business model consistency plays a vital role, as the internal fit of business model elements can generate reinforcing effects, thereby influencing performance and competitive advantage. Little research has been conducted to measure consistency and confirm these effects, especially within business model innovation. We tackle this issue by developing and testing a measurement of business model consistency, and investigate its relationship with business model innovation and its performance. We find evidence supporting the positive effect of consistency on innovation performance. We contribute to extant research by developing a concept and measurement for business model consistency based on contingency theory and empirically verifying it. Our findings underline that managers should pay close attention to the consistency of their business model designs during business model innovation.     

T–S fuzzy model identification based on a novel fuzzy c-regression model clustering algorithm

This paper proposes a novel approach for identification of Takagi–Sugeno (T–S) fuzzy model, which is based on a new fuzzy c-regression model (FCRM) clustering algorithm. The clustering prototype in fuzzy space partition is hyper-plane, so FCRM clustering technique is more suitable to be applied in premise parameters identification of T–S fuzzy model. A new FCRM clustering algorithm (NFCRMA) is presented, which is deduced from the fuzzy clustering objective function of FCRM with Lagrange multiplier rule, possessing integrative and concise structure. The proposed approach consists mainly of two steps: premise parameter identification and consequent parameter identification. The NFCRMA is utilized to partition the input–output data and identify the premise parameters, which can discover the real structure of the training data; on the other hand, orthogonal least square is exploited to identify the consequent parameters. Finally, some examples are given to verify the validity of the proposed modeling approach, and the results show the new approach is very efficient and of high accuracy.     

Connectivity and synchronization of Vicsek model

The collective behavior of multi-agent systems is an important studying point for the investigation of complex systems, and a basic model of multi-agent systems is the so called Vicsek model, which possesses some key features of complex systems, such as dynamic behavior, local interaction, changing neighborhood, etc. This model looks simple, but the nonlinearly coupled relationship makes the theoretical analysis quite complicated. Jadbabaie et al. analyzed the linearized heading equations in this model and showed that all agents will synchronize eventually, provided that the neighbor graphs associated with the agents＇ positions satisfy a certain connectivity condition. Much subsequent research effort has been devoted to the analysis of the Vicsek model since the publication of Jadbabaie＇s work. However, an unresolved key problem is when such a connectivity is satisfied. This paper given a sufficient condition to guarantee the synchronization of the Vicsek model, which is imposed on the model parameters only. Moreover, some counterexamples are given to show that the connectivity of the neighbor graphs is not sufficient for synchronization of the Vicsek model if the initial headings are allowed to be in [0,2π）, which reveals some fundamental differences between the Vicsek model and its linearized version.     

State-shared model for multiple-input multiple-output systems

This work proposes a method to construct a state-shared model for multiple-input multiple-output （MIMO） systems. A state-shared model is defined as a linear time invariant state-space structure that is driven by measurement signals-the plant outputs and the manipulated variables, but shared by different multiple input/output models. The genesis of the state-shared model is based on a particular reduced non minimal realization. Any such realization necessarily fulfills the requirement that the output of the state-shared model is an asymptotically correct estimate of the output of the plant, if the process model is selected appropriately. The approach is demomtrated on a nonlinear MIMO system - a physiological model of calcium fluxes that controls muscle contraction and relaxation in human cardiac myocytes.     

An image geometry model for METEOSAT†

METEOSAT is a spin-stabilized geostationary satellite which takes Earth images in three spectral channels, infrared, visible and water vapour absorption band. The time-dependent deviations of the satellite from nominal position, attitude and spin speed are described by an image geometry model. A new iterative procedure for refined attitude determination is used. The attitude is derived from northern and southern polar horizon scan limits which are extracted from infrared image data. The geometry model provides a deformation vector field which relates the ideal reference image to the actual image. The actual image is then mapped to the reference image by the nearest-neighbour rectification process.     

BSIM--making the first international standard MOSFET model

BSIM （Berkeley Short-Channel IGFET） became the first international industry standard model for simulation of MOS integrated circuits in 1997. The cumulative sales of ICs that have been designed with the aid of BSIM and produced for computing, communication, consumer, and industrial applications is estimated to be around 400 billion US dollars. From 0.35μm CMOS to multi-gate FinFET, BSIM serves a wide range of technologies. Many China educated researchers have contributed to its success.     

What makes a good process model?

There seems to be a never ending stream of new process modeling notations. Some of these notations are foundational and have been around for decades (e.g., Petri nets). Other notations are vendor specific, incremental, or are only popular for a short while. Discussions on the various competing notations concealed the more important question “What makes a good process model?”. Fortunately, large scale experiences with process mining allow us to address this question. Process mining techniques can be used to extract knowledge from event data, discover models, align logs and models, measure conformance, diagnose bottlenecks, and predict future events. Today’s processes leave many trails in data bases, audit trails, message logs, transaction logs, etc. Therefore, it makes sense to relate these event data to process models independent of their particular notation. Process models discovered based on the actual behavior tend to be very different from the process models made by humans. Moreover, conformance checking techniques often reveal important deviations between models and reality. The lessons that can be learned from process mining shed a new light on process model quality. This paper discusses the role of process models and lists seven problems related to process modeling. Based on our experiences in over 100 process mining projects, we discuss these problems. Moreover, we show that these problems can be addressed by exposing process models and modelers to event data.     

Graphical model construction based on evolutionary algorithms

1 Introduction Evolutionary algorithms(EAs) [1～5] are stochastic search and optimization techniques, which were inspired by the analogy of evolution and population genetics. They have been demonstrated to be effective and robust in searching very large, varied, spaces in a wide range of applications, including classification, machine learning, ecological, so- cial systems and so on. However, most of the common evo- lutionary algorithms using simple operators are incapable of learning the reg…     

Molecular dynamics simulation of nanofluid’s flow behaviors in the near-wall model and main flow model

The flow behaviors of nanofluids were studied in this paper using molecular dynamics (MD) simulation. Two MD simulation systems that are the near-wall model and main flow model were built. The nanofluid model consisted of one copper nanoparticle and liquid argon as base liquid. For the near-wall model, the nanoparticle that was very close to the wall would not move with the main flowing due to the overlap between the solid-like layer near the wall and the adsorbed layer around the nanoparticle, but it still had rotational motion. When the nanoparticle is far away from the wall (d > 11 Å), the nanoparticle not only had rotational motion, but also had translation. In the main flow model, the nanoparticle would rotate and translate besides main flowing. There was slip velocity between nanoparticles and liquid argon in both of the two simulation models. The flow behaviors of nanofluids exhibited obviously characteristics of two-phase flow. Because of the irregular motions of nanoparticles and the slip velocity between the two phases, the velocity fluctuation in nanofluids was enhanced.     

‘I-O-W’ model for national planning

The paper proposes an input-output-waste (I-O-W) model as an extension of the input-output model to incorporate waste management (WM) aspects in national planning. The waste aspects have not been explicitly considered in the existing literature on input-output economics. It is visualized that in the economic structure the waste parameter has a significant role to play, and the I-O-W model will be a closer representation of the economy than that of conventional input-output models. Any unnecessary input to or any undesirable output from the system is considered as waste. The basic framework of an I-O-W model is presented, which consists of a technology matrix, a final-demand matrix, and a WM policy matrix. The resource balance has been established by treating the input to be equal to the sum of output and waste. The technological coefficients are interlinked by wastivity indices. The projections of a hypothetical economy consisting of one production sector and one WM sector have been determined for material resources and for different WM policies.     

An application-semantics-based relaxed transaction model for internetware

An internetware application is composed by existing individual services, while transaction processing is a key mechanism to make the composition reliable. The existing research of transactional composite service (TCS) depends on the analysis to composition structure and exception handling mechanism in order to guarantee the relaxed atomicity. However, this approach cannot handle some application-specific requirements and causes lots of unnecessary failure recoveries or even aborts. In this paper, we propose a relaxed transaction model, including system mode, relaxed atomicity criterion, static checking algorithm and dynamic enforcement algorithm. Users are able to define different relaxed atomicity constraint for different TCS according to application-specific require- ments, including acceptable configurations and the preference order. The checking algo- rithm determines whether the constraint can be guaranteed to be satisfied. The enforce- ment algorithm monitors the execution and performs transaction management work ac- cording to the constraint. Compared to the existing work, our approach can handle com- plex application requirements, avoid unnecessary failure recoveries and perform the transaction management work automatically.     

Bootstrapping a robot’s kinematic model

We present a system that is able to autonomously build a 3D model of a robot’s hand, along with a kinematic model of the robot’s arm, beginning with very little information. The system starts by using exploratory motions to locate and centre the robot’s hand in the middle of its field of view, and then progressively builds the 3D and kinematic models. The system is flexible, and easy to integrate with different robots, because the model building process does not require any fiducial markers to be attached to the robot’s hand. To validate the models built by the system we perform a number of experiments. The results of the experiments demonstrate that the hand model built by the system can be tracked with a precision in the order of 1 mm, and that the kinematic model is accurate enough to reliably position the hand of the robot in camera space.