text stringlengths 128 2.05k |
|---|
Another critical point in analysis management is its convenience. The convenience and reproducibility should make a balance. A tool with high reproducibility but low convenience will force the users to bypass it. And thus the reproducibility will lose. |
People often overlook the significance of analysis management. They argue that management is just a bookkeeping work. However, management of the analysis is not an easy job. Recording the analysis details to an extent that others can repeat the analysis require quite a lot manpower. It is the present situation that mos... |
Why is analysis management difficult? One of the major sources of difficulty is from the preservation of data. Generally, data are sizable, and preserving the data using twice as many disks seems unwise. The codes and the data are usually lightly bound. The data are not generated intermediately after the codes are writ... |
and Data and codes are frequently modified or moved. For example, the shape of the fit functions can be adjusted. The upstream data may be modified and the whole analysis should rerun. |
Analyses in high energy physics are usually long-term, and they are usually highly complicated. Human beings are not reliable and analyzer often have terrible habits to organize the data and codes. The exact analysis step might be forgotten by the analyzer after a long time. |
The importance of analysis management has been recognized within high energy physics community. A conference preceding on data preservation |
is published in 2012. The CERN analysis preservation portal(CAP) is set up to prompt the systematic preservation of analysis for LHC experiments. CAP provides a centralized platform on which scientists can document their analysis as early as the start of a new project. CAP use REANA |
as backend, which is a platform for reusable research data analyses, receiving join effort from CERN IT, SIS, DASPOS and DIANA-HEP. |
In this paper, a proposal is raised to resolve the pain of analysis management. The proposal include an analysis architecture and a toolkit. The recent development of the container techniques both conceptually inspires the designation and technically provides a tool to realize the designation, as will be talked about i... |
Conceptual designation 2.1 Workflow architecture An analysis is always step by step. Each step requires some inputs and generates some outputs. If the inputs, outputs and steps are regarded as vertex, and directed edges are linked from the inputs to steps, as well as from steps to the outputs. The structure of an analy... |
Workflow exists in all analysis. The workflow of analysis contains fruitful information. It tells the procedure of the analysis, and it is also an invariant when the physical locations or names of the data or codes change. |
In the traditional way of analysis management, workflow is usually implicitly defined through the data location written in the codes, and it is usually ignored and seldom used. Workflow in analysis is also used in other domains like bioinformatics, medical imaging, astronomy, and chemistry. Workflow languages such as C... |
are developed. This explicit way of describing workflow has the advantage that the analysis procedure is transparent and directly readable to both the machine and the analyzer, clarifying the exact steps to do the analysis. So that the machine can run the analysis according to it, and analyzer can obtain knowledge of t... |
Additionally, there are two observations: A step in workflow does not necessarily know the whole workflow. It need only to know its predecessors to execute this specific step. |
The step can be abstract, i.e. only the topological structure of the analysis matters. Although workflow is widely used in management tool, it is seldom realized that the key function of workflow in analysis preservation is providing a tight binding between codes and data. |
The typical workflow concept has some slight differences with the one in this paper. In the workflow languages, they describe the exact steps to do the analysis. In this architecture, I choose to describe the workflow in a half-explicit way. It means that the workflow is not defined implicitly as described above. And i... |
Some of the analysis steps could be quite similar. They are better to be reused. I would like to separate the common part and the specialized part of an analysis step as “algorithm” and “task”. Loosely speaking, an “algorithm” contains the environment definition and the code template. An “algorithm” is will be used to ... |
It is also possible to refine the analysis architecture further. A task is always followed by some data. It is convenient to merge the “data” and “task” into one object. And the merge makes the binding of codes and data even tighter. |
The final architecture is illustrated in Fig. 2.2 Preservation and “impression” The code and the data in the above designation are still loosely bounded, since the modification on the codes would not impact the data immediately. There exists a conflict between the requirement of modification of codes and the preservati... |
2.2.1 Abstract and concrete layers Noticing the conflicts between different requirements discussed above, the analysis can be split into abstract layer and concrete layer according to the requirement of flexibility or reproducibility. To be specific, the abstract layer of an analysis is what the author write or operate... |
They can be executed to generate data. After the separation of the abstract layer and the concrete layer, the topological structure of an analysis may look like Fig. |
Since the abstract layer of the analysis contains only codes and some metadata, it has quite small size. They can be easily shared between analyzer to reuse and review. And the abstract layer can also be uploaded to a central portal for centralized management. |
2.2.2 “Impression” The reader may be confused about how such a splitting can balance the flexibility with reproducibility. In fact, it should move further to prompt “image” and “container” to “impression”. An “impression” is a version of an “image” or an “container”. It is able to generate data uniquely. And it is tigh... |
For example, as shown in Fig. , assume the original version of the analysis has four impressions, “Image 1.1”, “Image 2.1”, “Container 1.1”, and “Container 2.1”. After the modification of “Algorithm 2”, a new impression “Image 2.2” is created. At the same time, the “Task 2” will also generate a new impression “Containe... |
The abstract layer contains all the history of impression. Analysis can be reverted to a specific version if necessary. 2.3 Runner |
“Impression” is designed to determine the result. Conceptually, there should be a standalone runner to run the impression. The runner can be on a local machine and on a remote machine. The Analyzer sends “impression” to the runner. The runner executes the “impression” and generates result. And then the analyzer gets th... |
2.4 Status Each “task”, “algorithm” or “impression” can attach an status. The status is determined so that runner can manage the job sequence and analyzer can monitor the procedure of the analysis. |
2.4.1 Status of “image” and “container” Since an “impression” is immutable, it is not difficult to determine the status of an impression. If an “impression” is executed, the status of it will be “built” or “done” for “image” and “container” separately. The on going status is “building” or “running”. If it happened to b... |
2.4.2 Status of “task” and “algorithm” The status of algorithm and task can be “impressed” or “new”. One condition of “impressed” status is that the contents of an algorithm or task is the same as that of the corresponding impression. And the other condition is that dependences of the two parts also match. Otherwise, t... |
2.5 Global link Having the “impression” in hand, it is natural to link all the analysis in a collaboration together. An analyzer will not ask his/her colleague for data. The change of the upstream analysis will also be detected by the analyzer. |
Protocol 3.1 Introduction This protocol is designed to provide a standard way to organize the analysis in the high energy physics. If the analysis is organized in a standard way, the toolkit developer can develop software helping analyzer to manage it. The designation aims to validate conceptual designation of the anal... |
3.2 Organization 3.2.1 Chern repository All the analysis codes and metadata should be put in a directory. This directory is called a Chern repository. In the Chern repository, there exists a lot of subdirectories. Some of the directories contain hidden folder called “.chern”. Directory with the hidden “.chern” is calle... |
There are four types of object . They are algorithm task directory and project . Each object contains a file “README.md” used for documentation purpose. And a object also has a file named “.chern/config.json” serving as the configuration file of the object . The type of the object( algorithm task directory and project ... |
directory is just a folder to contain other sub object s. The object in a directory can be any type except directory project is exactly the same as the directory except that it is the root of the whole analysis repository. |
An algorithm contains the code templates and environment specification files(Dockerfile for temporatory), provided by collaboration. There is a folder called “.chern/impressions” containing all the impressions corresponding to the algorithm |
In a task , there should be a file “parameter.json” containing the parameters of the task. There is also a folder called “.chern/impressions” as algorithm |
The connection between the task s and algorithm s, i.e. workflow, is specified in the configuration files of the task s and algorithm s. In the configuration of a task , the algorithm and task s that this task depends on are recorded, in the form of the relative path to the project . The depending task has also an “ali... |
Other files or directories that are not organized as above statements, do not belong to the repository and should be ignored. 3.2.2 Impression |
An impression is a version of a task or algorithm . It has a unique id and is physically stored in the “.chern/impressions/(ID)” under the task or algorithm , where the (ID) refer to the id of the impression . Under the “.chern/impressions/(ID)”, there exists a file named “config.json”, called the configuration file of... |
The above information stored in a impression is enough for running the impression anywhere. 3.2.3 Parameter transfer and I/O map |
The parameters defined in a task should be able to be transfered to the corresponding algorithm Let us take an ROOT or C++ application as example. When the analyzer writes a template code in ROOT for an algorithm he/she could include a header file called “arguments”. After including the header, he/she can use “paramete... |
Dealing with the inputs and outputs is similar. As discussed above, each predecessor of the task has an alias. After including the header, analyzer can use “folder[(ALIAS)]” as a std::string type value. This value means the storage location of the data of the predecessor task . The output directory to store the result ... |
The runner should generate the header file “arguments”. Other programming languages can be supported in a similar way. However, this do not mean that the standard transfer header should be created for all the programming language that we can image. Creating “header file” for script language like bash, tcsh or python, a... |
3.2.4 Pure data task It is a common case that the starting point of an data analysis is external data. However, as the “data” can only exists in a “task”, there is no place to put the external data. This can be solved in the following way. The Analyzer creates a file named “data.json” under a task and pretend that it c... |
3.3 Operations The operation applied to the repository should be restricted in order to keep the structure of the operation. The following operations are allowed to use and should be regarded as the atomic operations. |
Creating repository: making an empty folder and creating file “.chern/config.json” and “.chern/project.json” under it. Writing the object type of “project” into file “.chern/config.json”. |
Creating directory : making an empty folder and creating file “.chern/config.json” under it. Writing the object type of “directory” into file “.chern/config.json”. |
Creating algorithm : making an empty folder and creating file “.chern/config.json” under it. Writing the object type of “algorithm” into file “.chern/config.json”. |
Creating task algorithm : make an empty folder and creating file “.chern/config.json” under it. Writing the object type of “task” into file “.chern/config.json”. |
Link an algorithm and a task : adding the task to the successor list of the algorithm in the configuration file of the algorithm . Adding the algorithm to the predecessor list of the task |
Unlink and algorithm and a task : removing the task from the successor list of algorithm . Removing the algorithm from the predecessor list of the task |
Link a task to a second task : adding the second task to the successor list of the first task . Adding the first task to the predecessor list. |
Unlink the relation of a task to a second task removing the second task of the successor list of the first task . Removing the first task from the predecessor list of the second task . Removing the alias. |
Moving object : moving the directory. Since the object recorded in its predecessors and successors in the form of the path relative to the root of the whole repository. The configuration files of its predecessors and successors should be modified to use the new relative path. |
Copying object : copying the directory. Unlink the predecessors and the successors. Removing object : unlink the predecessors and the successors. Removing the directory. |
Modify “README.md”: just editing “README.md”. Creating impression : Once an impression is required to be created for an algorithm or a task , firstly, the impression of the preceding task or algorithm is checked to be available. After that, the contents of the task or algorithm and the preceding impression define the r... |
Technical designation and prototype A prototype toolkit named “Chern” is designed to realize the conceptual designation. The github repository of the toolkit is on |
. And the documents for user can be found in A few examples including some toy examples and some real analyses can be found on github. |
4.1 Frontend A toolkit should develop commands to realize the operation defined in the above section and their combination helping the analyzer to operate the Chern repository conveniently. I choose to use IPython notebook to hide the details of managing Chern repository as much as possible. The commands such as “mv”, ... |
4.2 Chern Machine 4.2.1 Execution of impression Impression should be able to be executed on any machine to generate the exact same results. In the past, to achieve such a requirement, a virtual machine is needed. Nowadays, the development of container technique makes such a requirement much easier to achieve. The runne... |
4.2.2 Communication The communication protocols between the Chern repository and the runner, and between runners are not defined in the protocol above, since different communication methods are in principle equivalent and it is better to support many of them. In the current prototype, I only use http for the communicat... |
Discussion and Future development 5.1 Use cases and advantages In the following cases, using the proposed Chern architecture will be convenience. They often occur in the analysis of high energy physics. |
Using the management system, analyzers can directly read the analysis steps from the repository. Analysis steps are preserved and for every data. The detailed steps to generate data are clear. |
The name and location of the directory are not useful for backend to run the analysis. Therefore, they can be arbitrary modified without changing any result. When adding new parallel steps to analysis, this feature is especially useful. |
Similar analysis can easily be constructed. An analyzer can just copy a Chern repository from others and perform some slight tunning to create a new analysis. |
5.2 Future development 5.2.1 Time machine Although the full history is recorded in the Chern repository. The time machine function is not realized in the prototype. Reverting the analysis is often useful and this function will be added in the future version of the toolkit. |
5.2.2 Monitoring A job monitoring system can be developed to manage the job status and disk usage. 5.2.3 Frontend The current frontend of the toolkit is based on IPython notebook. But the architecture is in fact independent of the analysis. The frontend can also be web application or GUI application. They might be more... |
5.2.4 Backend Technically, the current prototype of Chern only naively run, while RENAN support a lot of advanced techniques. REANA uses container technologies(Docker) and aims at supporting several different declarative workflow engines(Yadage, CWL), several different computing cloud back-ends for job execution(Kubert... |
Conclusion In this paper, an architecture to manage an analysis is proposed. With the new architecture, the workflow is constructed in an apparent way without losing convenience. The requirement of preservation is also satisfied. A novel toolkit is provided for analyzers. So that they can do the analysis as what they d... |
The future development of the architecture and the toolkit is also discussed. The standard, time machine function, the new frontend and the new backend are also proposed. |
Acknowledgements I would like to express my acknowledgements to Sidan Chen, Mengzhen Wang, Li Xu, and Zhenwei Yang who help to polish this paper. |
# Source: arxiv 1806.08808 # Title: Complexity Matching and Requisite Variety # Sections: all # Downloaded: 2026-03-02T08:57:31.396703+00:00 |
Complexity Matching and Requisite Variety Korosh Mahmoodi 1,* Bruce J. West Paolo Grigolini Department of Social and Decision Sciences, Carnegie Mellon University, Pittsburgh, PA 15213, USA |
Information Science Directorate, Army Research Office, Research Triangle Park, NC 27708, USA Center for Nonlinear Science, University of North Texas, P.O. Box 311427, Denton, TX 76203, USA |
* corresponding author, koroshm@andrew.cmu.edu Abstract Complexity matching characterizes the role of information in interactions between systems and can be traced back to the 1957 Introduction to Cybernetics by Ross Ashby. We argue that complexity can be expressed in terms of crucial events, which are self-organizatio... |
Author summary This paper is devoted to the control of complex dynamical systems, inspired by real processes of biological and sociological interest. The concept of complexity we adopt focuses on the assumption that the processes of self-organization generate intermittent fluctuations and that the time interval between... |
INTRODUCTION The mathematician Norbert Wiener observed in a 1948 ‘popular’ lecture that the complex networks in the social and life sciences appear to behave differently from, but not in contradiction to, the laws in the physical sciences. He makes the point that the force laws and therefore control of social phenomena... |
to the scientific community in which his interest in control and communication within and between animals and machines was made clear. |
The economic webs of global finance and stock markets; the social meshes of governments and terrorist organizations; the transportation networks of planes and highways; the ecowebs of food networks and species diversity; the physical wicker of the Internet; the bionet of gene regulation, and so on, have all been modele... |
West et al. , among many others, noted that complexity often arises in a network when the power spectrum [MATH] takes on an IPL shape: |
[EQUATION] with the IPL index [MATH] in the interval [MATH] . In fact, this [MATH] variability is taken by many scientists to be the signature of complexity and appears in a vast array of phenomena including the human brain |
, body movements , music , physiology , genomics and sociology . They then used the IPL index as a measure of a network’s complexity and reviewed the literature arguing that two complex interacting networks exchange information most efficiently when the IPL indices of the two networks match. The hypothesis of the |
complexity management effect form of the WR was proven by Aquino et al. using averages over PDFs and more generally using time averages |
In less than a decade after the initial formulation of cybernetics, Ross Ashby, captured in his introduction to the subject , the difficulty of regulating biological systems and that “the main cause of difficulty is the variety in the disturbances that must be regulated against”. This insightful observation, which was ... |
requisite variety with the more recent term complexity matching effect (CME) . The term complexity matching has been widely used in the recent past to include such information exchange activities as side-by-side walking |
, ergometer rowing , syncopated finger tapping , dyadic conversation , and interpersonel coordination . These synchronizations are today’s realizations of the regulation of the brain, in conformity with the observations of Ashby. |
Crucial events and the brain: For the purposes of the present paper it is important to stress that there exists further research directed toward the foundations of social learning |
that is even more closely connected to Ashby’s challenge of the regulator and the regulated sharing a common level of complexity. In fact, this social learning research aims at evaluating the transfer of information from the brain of one player to that of another by way of the interaction the two players established th... |
, which is to say, it is a virtual reality experiment capturing the social cognition shared by a pair of humans. The results are exciting in that the trajectories of pairs of players turn out to be significantly synchronized. But even more important than synchronization is the fact that the trajectories of the two avat... |
Herein we provide a theoretical rational for the universal structure representing the brains of the pairs of interacting individuals, based on the CME. More broadly the theory can be adapted to the communication, or information transfer, between the heart and the brain |
of a single individual. The key to this understanding is the existence of crucial events. In a system as complex as the human brain |
there is experimental evidence for the existence of crucial events, which, for our purposes here, can be interpreted as organization rearrangements, or renewal failures. The time interval between consecutive crucial events is described by a waiting-time IPL PDF |
[EQUATION] The crucial events generate ergodicity breaking and are widely studied to reveal fundamental biological statistical properties |
The transfer of information between interacting systems has been addressed using different theoretical tools, examples of which include: chaos syncronization |
, self-organization , and resonance . However, none of these theoretical approaches have to date explained the experimental results that exist for the correlation between the dynamics of two distinct physiological systems |
. Herein we relate this correlation to the occurrence of crucial events, which are responsible for the generation of [MATH] variability with an IPL spectrum having an IPL index [MATH] and for the results of a number of psychological experiments including those of Correll |
. The experimental data imply that activating cognition has the effect of making the IPL index [MATH] cross the barrier between the Lévy and Gauss basins of attraction, namely making [MATH] |
. This is in line with Heidegger’s phenomenology The crossing of a basin’s boundary is a manifestation of the significant effect of violating the linear response condition, according to which a perturbation should be sufficiently weak that it does not affect a system’s dynamic complexity |
. The experimental observation obliged us to go beyond the linear response theory adopted in earlier works in order to explain the transfer of information from one complex system to another and to formally prove the WR. This information transfer was accomplished through the matching of the IPL index of the crucial even... |
. This is consistent with the general idea of the CME, with the main limitation being that the perturbation intensity is sufficiently small that it is possible to observe the influence of the perturbing system on the perturbed system through ensemble averages, namely a mean over many realizations |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.