Abstract
The article analyses the possibility of applying immune-like processing of legal texts in the legislative process. It also discusses the required format for recording legal information and the relationships between the formats for recording legal information and the methods used to analyse it. On this ground we argue that the formal analysis of a legal text must extend beyond its logical consistency. This is so due to the fact that the quality of legislation depends not only on the lack of internal contradictions within a legal text but also on the number of requirements a legal text has to meet. Legislation should not only be consistent but also coherent, uniform and comprehensible. This requires, inter alia, consistent terminology, a lack of redundancy, appropriate references and compliance with the requirements set forth in the principles of legislation . In order to achieve the above-mentioned aims legislators are beginning to apply new tools to complement traditional approaches. As a consequence, legal drafting is nowadays aided by various information and communication technologies usually based on classic editing tools and algorithms for text processing. However, due to deficiencies in already existing solutions it is important to consider using novel, adaptive and specially dedicated algorithms, which would allow for similarity analyses of legal texts and in particular would detect patterns, relationships or coincidences in their content and structure. Such algorithms can be based on computational intelligence, and more specifically on artificial immune systems . As a consequence, the article will discuss both the scope of application and the capacity of tools based upon the idea of natural immune systems as means of increasing the quality of legislation .
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
Cyrul et al. (2014).
- 2.
Sartor et al. (2011).
- 3.
Francesconi (2011).
- 4.
XML documents contain mark-ups which encode a description of the document’s storage layout and logical structure. XML makes it possible to impose constraints on the storage layout and logical structure. See: W3C Recommendation, February 10, 1998. http://www.w3.org/TR/1998/REC-xml-19980210. Accessed 16 December 2013.
- 5.
Language for expressing constraints about XML documents. For example, it can be used to provide a list of elements and attributes in a vocabulary or to place restrictions on when elements and attributes can appear. See: http://www.w3.org/standards/xml/schema. Accessed 16 December 12.
- 6.
See more http://www.digitales.oesterreich.gv.at/site/6565/default.aspx. Accessed 6.05.2014.
- 7.
See more https://www.ministerialtidende.dk/Forms/L0500.aspx?page=5. Accessed 6.05.2014.
- 8.
See more http://www.metalex.nl/, http://www.metalex.eu/. Accessed 6.05.2014.
- 9.
See more http://www.estrellaproject.org/doc/D3.1-General-XML-formats-For-Legal-Sources.pdf. Accessed 6.05.2014.
- 10.
See more http://webarchive.nationalarchives.gov.uk/20100225080346/opsi.gov.uk/legislation/schema/. Accessed 6.05.2014.
- 11.
See more http://www.ittig.cnr.it/Ricerca/UnitaEng.php?Id=40. Accessed 6.05.2014.
- 12.
See more http://www.svri.ch/de/pdf/CHLexMLReference1.0.pdf. Accessed 6.05.2014.
- 13.
See more http://www.thelaw.tas.gov.au/about/enact.w3p. Accessed 6.05.2014.
- 14.
See more http://www.agls.gov.au/. Accessed 6.05.2014.
- 15.
See more in Palmirani and Vitali (2011).
- 16.
See: http://formex.publications.europa.eu/index.html. Accessed 16 December 2013.
- 17.
For example, Legalis (http://www.legalis.pl/wydarzenia/dokument/rewolucja-w-legislacji/), EDAP (http://mac.bip.gov.pl/elektroniczna-forma-aktow-prawnych/elektroniczna-forma-aktow-prawnych.html), xmLeges (http://www.xmleges.org/eng/), DALOS (http://www.dalosproject.eu/). Accessed 6.05.2014.
- 18.
As described by Cyrul et al. (2014).
- 19.
Note that text or numerical values entered by the user can not conform to expected constraints, patterns, data types (including data entered instead of numeric, incorrect values/grammatical form).
- 20.
Analysed data can be incompatible with the expected pattern, schema, template, law, code etc.
- 21.
A large distance indicates little similarity.
- 22.
The natural immune system protects bodies from pathogens (infectious agents). Antigens make it possible to identify invading agents by the immune cells and molecules to provoke an immune response. In general, two types of cells are considered: lymphocytes T (called as T-lymphocytes, T-cells) and B (B-lymphocytes, B-cells), which are constantly produced by the human body. When T-cells are used to identify a new kind of pathogen through binding, the primary immune response is trigged. A number of antibodies (cells of type B) are produced to eliminate an antigen from the body and keep this knowledge in the immune memory. Immune memory makes it possible to perform secondary immune response when a similar antigen is encountered again (Timmis et al. 2004; Somayaji et al. 1998; Pełech and Duda 2005).
- 23.
Clonal selection mechanism makes it possible to adapt the system by producing a number of antigens in response to a pathogen. Activated lymphocytes are widely cloned and modified to match with and adapt to the pathogen. They are then evaluated and selected to leave a population of well-adapted lymphocytes in the system.
- 24.
The efficiency of pathogen detection increases in tandem with the increasing frequency with which the same antigen is encountered in time (i.a. a number of clonal selections are triggered).
- 25.
Somatic hyper mutation (SHM) allows for the diversification of lymphocytes to detect (recognize) unknown cells (objects, structures) by producing a number of diversified structures. As a result, well-matched antibodies are generated and the immune system is adapted. To maintain the diversity of detectors genetic operators are employed, including mutation (typically, a change in selected value(s)) and crossover (a new detector is produced through its combination with another detector).
- 26.
During the maturation process, self-reactive lymphocytes are removed from the system. Thus, the process of nonself detection is based on the recognition of unknown or untypical cells/object/structures.
- 27.
When a pathogen is encountered for the first time, the first immune response is triggered and the pathogen is removed (mutation and clonal selection are performed). The system retains information on the detection scheme in the immune memory used during subsequent recognitions of the pathogen. This memory is adaptive (the current activities of the system cells are retained; efficient schema are strengthened). Note that the accuracy of detection and the possibility of rapid removal from the system is also due to affinity maturation and thus clonal selection.
- 28.
A cell which represents information on a book can be encoded as text and numerical data on inter alia the author(s), title, major form (novel, poem etc.), genre (tragedy, comedy, epic, lyric etc.), technique (prose, poetry), physical format (paper, electronic, hardcover), International Standard Book Number (ISBN), release date, publisher, indexes in catalogues (databases). In such a case, each book should be encoded with the same encoding schema. Self-cells can represent, for example, the correct information on a book.
- 29.
To retain information on text-based content as a cell, it is important to encode both text component and related attributes.
- 30.
More precisely: recognition of cells other than self-cells.
- 31.
For example, improve an analysed text, remove incorrect words, tagging.
- 32.
Detectors can be adjusted to analyse content, consistency or structure.
- 33.
A typical parameter of an adaptive algorithm is window width. For example, the same algorithm can be used to analyse one sentence (a narrow window width) and a whole paragraph (a wide width).
- 34.
To achieve this result cells and detectors are adapted. This can be done in a genetic way.
- 35.
Parallel systems can compute many computational tasks simultaneously.
- 36.
Computations can be distributed on different tasks.
- 37.
Adaptive systems are based on algorithms which can adapt to new data (information) received, processing conditions etc.
- 38.
An artificial immune system may consist of subsystems (modules) based on supervised learning algorithms aimed at producing the inferred function on the basis of training data prepared by the user.
- 39.
See footnote 15 and Cyrul et al. (2014).
- 40.
Data stored in the system log on user activity can include information on the following: login session, names opened/edited documents (and related time), entered/deleted words/signs, additional file usage and printing. Such services can be supported by immune-inspired systems by, for example, analysing untypical user activity.
- 41.
Dedicated immune-like algorithms can provide advanced monitoring of changes in documents (normative acts) edited by the legislator, including punctuation marks, word order, untypical features, the time and intensity of changes and the related user’s names.
- 42.
A large set of available legal documents includes normative acts that are binding on a normative act drafted by the legislator. A system used for supporting legislation should accurately capture incorrect references, definitions etc. in real time.
- 43.
For example, spectral analysis (Fourier analysis) used with time series allows for their representation as sines and cosines, thereby signalling decomposition into fast-variable and slow-variable components. In particular, it enables periodicity analysis.
- 44.
Noise reduction refers to the removal of useless (insignificant) information from the input data that impede reliable analysis. Considering the time series, trend extraction from data may allow for more accurate forecasts. In the case of text-based content, individual letters or punctuation marks can be statistically irrelevant for text analyses.
- 45.
The aim of unifying and standardising a diversified input data set is to enable the use of numerical procedures, calculations, transformations of the same type and to make information storage, search and retrieval.
- 46.
The goal is to obtain a homogenous, uniform, disambiguate data set which doesn’t contain unimportant variables.
- 47.
In many cases it is reasonable or necessary to increase or decrease the weight of attributes, specific data or dependencies. For example, small changes/differences between values can be enhanced by exponentiation, the importance of selected words can be weakened etc.
- 48.
The purpose of data processing and related assumptions imply the number of attributes and their complexity.
- 49.
According to the principle of negative selection, the detection process is based on pathogens (non-selfs) being recognized by lymphocytes.
- 50.
Such an analysis cay support a system or application for built-in data protection mechanisms (providing data authentication, confidentiality, integrity, availability).
- 51.
Feature space refers to the n−dimensions where variables (analysed objects) are considered. Variables are viewed as features.
- 52.
The Jaccard similarity coefficient is used for similarity analysis of sets A and B. For binary data it is calculated as the length of intersection of two input sets (A ∩ B) divided by the length of the union of the sets (A ∪ B). The Jaccard distance is calculated by subtracting the coefficient from 1.
- 53.
Cosine similarity is expressed as the normalized dot product of sets A and B (for text-based content matching: term frequency vectors): the dot product of A and B divided by the length of A multiplied by the length of B. For vector representation of documents, cosine similarity represents the cosine of the angle between documents. The value of the cosine distance equal to 1 indicates the same documents while a value equal to zero indicates no relationship between the documents.
- 54.
- 55.
Non-efficient detectors are to be removed from the system through their deactivation, by changing the structure or by fixing the parameters.
- 56.
The adaptation process must be repeated until there are no detectors binding the self-structures.
- 57.
Pathogen – referring to the immune paradigm.
- 58.
Although in such a case the detection scheme formulated above can be regarded as adequate so as to avoid examining system cells by brute-force methods as well as to utilize efficient and verified detection schemas (profiles), exploiting the system memory in the detection process seems to be valuable.
- 59.
To achieve a high level of accuracy and sensitivity in text processing, in many cases it may be necessary to perform time-consuming calculations or to ask the user (legislator) to confirm changes manually.
- 60.
A complementary detector set may consist of detectors designed to analyze specific, particular features of a text.
- 61.
Art. 54 Pożytkami prawa sa̧ dochody, które prawo to przynosi zgodnie ze swym społeczno-gospodarczym przeznaczeniem. (Civil Code, Journal of Laws of 1964, no. 16, item 93 with amendments.) Article 54. Proceeds which a right produces according to its social and economic purpose shall be profits from that right.
- 62.
Directly encoding a date/time requires further data categorization or precise formulation of database queries.
- 63.
The methods discussed here and used for similarity analyses are applied in the field of text analysis relatively rarely.
- 64.
Calculated as the square root of the sum of the squares of the differences between corresponding points/vectors.
- 65.
Calculated as the greatest difference between points/vectors corresponding points/vectors.
- 66.
Art. 53. §1. Pożytkami naturalnymi rzeczy sa̧ jej płody i inne odła̧czone od niej czȩści składowe, o ile według zasad prawidłowej gospodarki stanowia̧ normalny dochód z rzeczy. §2. Pożytkami cywilnymi rzeczy sa̧ dochody, które rzecz przynosi na podstawie stosunku prawnego. Ustawa z dnia 23 kwietnia 1964 r. – Kodeks cywilny (Journal of Laws of 1964, no. 16, item 93 with amendments.) Article 53. §1. A thing’s produce and other component parts detached from it, as long as according to the principles of careful economic management they constitute the usual proceeds from the thing, shall be natural profits from the thing. §2. Proceeds which the thing produces on the basis of a legal relation shall be civil profits from the thing.
- 67.
Art. 54 Pożytkami prawa sa̧ dochody, które prawo to przynosi zgodnie ze swym społeczno-gospodarczym przeznaczeniem. (Civil Code, Journal of Laws of 1964, no. 16, item 93 with amendments.) Article 54. Proceeds which a right produces according to its social and economic purpose shall be profits from that right.
- 68.
Art. 55. §1. Uprawnionemu do pobierania pożytków przypadaja̧ pożytki naturalne, które zostały odła̧czone od rzeczy w czasie trwania jego uprawnienia, a pożytki cywilne - w stosunku do czasu trwania tego uprawnienia. §2. Jeżeli uprawniony do pobierania pożytków poczynił nakłady w celu uzyskania pożytków, które przypadły innej osobie, należy mu siȩ od niej wynagrodzenie za te nakłady. Wynagrodzenie nie może przenosić wartości pożytków. Ustawa z dnia 23 kwietnia 1964 r. – Kodeks cywilny (Civil Code, Journal of Laws of 1964, no. 16, item 93 with amendments.) Article 55. §1. A person entitled to collect profits shall collect these natural profits which have been detached from the thing during his entitlement, and civil profits - in proportion to the period of this entitlement’s duration. §2. If the person entitled to collect profits made expenses aimed at obtaining profits which have fallen to another person, he shall be entitled to the remuneration for these expenditures. The remuneration may not exceed the value of the profits.
- 69.
Art. 993. Przy obliczaniu zachowku nie uwzglȩdnia siȩ zapisów zwykłych i poleceń, natomiast dolicza siȩ do spadku, stosownie do przepisów poniższych, darowizny oraz zapisy windykacyjne dokonane przez spadkodawcȩ. Ustawa z dnia 23 kwietnia 1964 r. – Kodeks cywilny (Civil Code, Journal of Laws of 1964, no. 16, item 93 with amendments.) Article 993. When calculating the reserved portion, ordinary legacies and instructions shall not be taken into account, unlike donations and specific bequests made by the decedent, which shall be added to the estate, pursuant to the below-mentioned provisions).
- 70.
Art. 796. Jeżeli przepisy tytułu niniejszego albo przepisy szczególne nie stanowia̧ inaczej, do umowy spedycji stosuje siȩ odpowiednio przepisy o umowie zlecenia. Ustawa z dnia 23 kwietnia 1964 r. – Kodeks cywilny (Civil Code, Journal of Laws of 1964, no. 16, item 93 with amendments.) Article 796. If the provisions of the present title or specific provisions do not provide otherwise, the provisions on on the contract of mandate shall apply accordingly.
- 71.
To achieve relatively high detection accuracy a detection algorithm should be suitable for input data properties, detailed data processing purposes, the intended use of the processing results etc.
References
Ademokun, Alexander A., and Deborah Dunn-Walters. 2010. Immune responses: Primary and secondary. In eLS, Wiley Oline Library. http://onlinelibrary.wiley.com/doi/10.1002/9780470015902.a0000947.pub2/full.
Ashby, F. Gregory, and Nancy A. Perrin. 1988. Toward a unified theory of similarity and recognition. Psychological Review 95:124–150.
Charikar, Moses S. 2002. Similarity estimation techniques from rounding algorithms. In Proceedings of the thiry-fourth annual ACM symposium on theory of computing, ed. John Reif, 380–388. New York: ACM.
Cyrul, Wojciech, Jerzy Duda, Janusz Opiła, and Tomasz Pełech-Pilichowski. 2014. Informatyzacja tekstu prawa. Perspektywy zastosowania jȩzyków znacznikowych. Warsaw: Wolters Kluwer SA.
Dasgupta, Dipankar. 2006. Advances in artificial immune systems. IEEE Computational Intelligence Magazine 1(4):40–49.
de Castro, Leonardo N., and Jonathan Timmis. 2002. Artificial immune systems: A new computational approach. London: Springer.
Deerwester, Scott, Susan T. Dumais, George W. Furnas, Thomas K. Landauer, and Richard Harshman. 1990. Indexing by latent semantic analysis. Journal of the American Society of Information Science 41(6):391–407.
Forrest, Stephanie, Alan S. Perelson, Lawrence Allen, and Rajesh Cherukuri. 1994. Self-nonself discrimination in a computer. In Proceedings of the 1994 IEEE symposium on security and privacy, 202. Washington, DC: IEEE.
Francesconi, Enrico. 2011. A review of systems and projects: Management of legislative resources. In Legislative XML for the semantic Web, ed. Giovanni Sartor, Monica Palmirani, Enrico Francesconi, and Maria A. Biasiotti, 173–188. Berlin/Heidelberg: Springer.
Glickman, Matthew, Justin Balthrop, and Stephanie Forrest. 2005. A machine learning evaluation of an artificial immune system. Evolutionary Computation Journal 13(2):179–212.
Gonzalez, Fabio, Dipankar Dasgupta, and Robert Kozma. 2002. Combining negative selection and classification techniques for anomaly detection. In Proceedings of the congress on evolutionary computation, (1), 705–710. Honolulu: IEEE.
Hand, David, Heikki Mannila, and Padhraic Smyth. 2001. Principles of data mining. Cambridge: MIT Press.
Hunt, John E., and Denise E. Cooke. 1996. Learning using an artificial immune system. Journal of Network and Computer Applications 19:189–212.
Jolliffe, Ian T. 2002. Principal component analysis. New York, Berlin, Heidelberg: Springer-Verlag.
Karakose, Mahmet. 2013. Reinforcement learning based artificial immune classifier. The Scientific World Journal. Hindawi. http://downloads.hindawi.com/journals/tswj/2013/581846.pdf. Accessed 15 Jan 2014.
Kumar, Akshat, and Shivashankar Nair. 2007. An artificial immune system based approach for English grammar checking. In Artificial immune systems, ed. de Castro and Leonardo N., et al. LNCS, vol. 4628, 348–357. Berlin/Heidelberg: Springer.
Landauer, Thomas K., and Susan T. Dumais. 1997. A solution to Plato’s problem: The latent semantic analysis theory of acquisition. Psychological Review 104(2):211–240.
Mahmoud, Tarek M., and Ahmed M. Mahfouz. 2012. SMS spam filtering technique based on artificial immune system. IJCSI International Journal of Computer Science Issues 9(2):589–597.
Palmirani, Monica, and Fabio Vitali. 2011. Akoma-Ntoso for legal documents. In Legislative XML for the semantic Web, ed. Giovanni Sartor, Monica Palmirani, Enrico Francesconi, and Maria A. Biasiotti, 75–100. Berlin/Heidelberg: Springer.
Pełech, Tomasz, Jan T. Duda. 2005. Immune algorithm of stock rates parallel monitoring. In Information systems and computational methods in management, ed. Jan T. Duda J.T, 183–193. Krakow: AGH-UST Publishing.
Pełech-Pilichowski, Tomasz, Wojciech Cyrul, and Piotr Potiopa. 2014. On problems of automatic legal texts processing and information acquiring from normative acts. In Advances in business ICT, ed. Maria Mach-Król, and Tomasz Pełech-Pilichowski, 53–67. Cham: Springer.
Romero, Andres, Fernando Nino. 2007. Keyword extraction using an artificial immune system. In: Proceedings of the 9th annual conference on Genetic and evolutionary computation, 181. London: ACM.
Samsudin, Norlela, Mazidah Puteh, Abdul R. Hamdan, and Mohd Z.A. Nazri. 2012. Is artificial immune system suitable for opinion mining? In: 4th conference on data mining and optimization, Langkawi, 131–136.
Sartor, Giovanni, Palmirani Monica, Francesconi Enrico, and Maria Angela Biasiotti. eds. 2011. Legislative XML for the semantic Web. Dordrecht: Springer.
Schadwinkel, Stefan, and Werner Dilger. 2006. A dynamic approach to artificial immune systems utilizing neural networks. In Proceedings of the 8th annual conference on genetic and evolutionary computation, 131–132. New York: ACM.
Somayaji, Anil, Steven Hofmeyr, and Stephanie Forrest. 1998. Principles of a computer immune system. In Proceedings of the 1997 workshop on new security paradigms, 75–82. New York: ACM.
Tang, Na, and V. Rao Velmuri. 2005. An artificial immune system to document clustering. In Proceedings of the 2005 ACM symposium on applied computing, ed. Hisham Haddad, Lorie M. Liebrock, Andrea Omicini, and Roger L. Wainwright, 918–922. New York: ACM.
Timmis, Jonathan, Mark Neal, and John Hunt. 2000. An artificial immune system for data analysis. Biosystems 55(1):143–150.
Timmis, Jonathan, Thomas Knight, Leonardo N. de Castro, and Emma Hart. 2004. An overview of artificial immune systems. In Computation in cells and tissues: Perspectives and tools for thought, ed. Ray Paton, Hamid Bolouri, Mike Holcombe, J. Howard Parish, and Richard Tateson, 51–86. Berlin, Heidelberg, New York: Springer Verlag.
Wierzchoń, Sławomir T. 2001. Artificial immune systems. Theory and applications. Warsaw, Poland: Akademicka Oficyna Wydawnicza EXIT.
Wolfram, Stephen. 1986. Approaches to complexity engineering. Physica D (22):385–399.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2015 Springer International Publishing Switzerland
About this chapter
Cite this chapter
Pełech-Pilichowski, T., Cyrul, W. (2015). Computer-Aided Legislation Based on Immune-Like Processing of Legal Texts. In: Araszkiewicz, M., Płeszka, K. (eds) Logic in the Theory and Practice of Lawmaking. Legisprudence Library, vol 2. Springer, Cham. https://doi.org/10.1007/978-3-319-19575-9_20
Download citation
DOI: https://doi.org/10.1007/978-3-319-19575-9_20
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-19574-2
Online ISBN: 978-3-319-19575-9
eBook Packages: Humanities, Social Sciences and LawLaw and Criminology (R0)