Jouko Vankka

Localization Services for Online Common Operational Picture and Situation Awareness

Many operations, be they military, police, rescue, or other field operations, require localization services and online situation awareness to make them effective. Questions such as how many people are inside a building and their locations are essential. In this paper, an online localization and situation awareness system is presented, called Mobile Urban Situation Awareness System (MUSAS), for gathering and maintaining localization information, to form a common operational picture. The MUSAS provides multiple localization services, as well as visualization of other sensor data, in a common frame of reference. The information and common operational picture of the system is conveyed to all parties involved in the operation, the field team, and people in the command post. In this paper, a general system architecture for enabling localization based situation awareness is designed and the MUSAS system solution is presented. The developed subsystem components and forming of the common operational picture are summarized, and the future potential of the system for various scenarios is discussed. In the demonstration, the MUSAS is deployed to an unknown building, in an ad hoc fashion, to provide situation awareness in an urban indoor military operation.

Instant message classification in Finnish cyber security themed free-form discussion

Instant messaging enables rapid collaboration between professionals during cyber security incidents. However, monitoring discussion manually becomes challenging as the number of communication channels increases. Failure to identify relevant information from the free-form instant messages may lead to reduced situational awareness. In this paper, the problem was approached by developing a framework for classification of instant message topics of cyber security-themed discussion in Finnish. The program utilizes open source software components in morphological analysis, and subsequently converts the messages into Bag-of-Words representations before classifying them into predetermined incident categories. We compared support vector machines (SVM), multinomial naïve Bayes, and complement naïve Bayes (CNB) classification methods with five-fold cross-validation. A combination of SVM and CNB achieved classification accuracy of over 85 %, while multiclass SVM achieved 87 % accuracy. The implemented program recognizes cyber security-related messages in IRC chat rooms and categorizes them accordingly.

Vector Space Representations of Documents in Classifying Finnish Social Media Texts

Computational analysis of linguistic data requires that texts are transformed into numeric representations. The aim of this research is to evaluate different methods for building vector representations of text documents from social media. The methods are compared in respect to their performance in a classification task. Namely, traditional count-based term frequency-inverse document frequency (TFIDF) is compared to the semantic distributed word embedding representations. Unlike previous research, we investigate document representations in the context of morphologically rich Finnish. Based on the results, we suggest a framework for building vector space representations of texts in social media, applicable to language technologies for morphologically rich languages. In the current study, lemmatization of tokens increased classification accuracy, while lexical filtering generally hindered performance. Finally, we report that distributed embeddings and TFIDF perform at comparable levels with our data.

Situation awareness requirements for a critical infrastructure monitoring operator

This paper presents a set of situation awareness (SA) requirements for an operator who monitors critical infrastructure (CI). The requirements consist of pieces of information that the operator needs in order to be successful in their work. The purpose of this research was to define a common requirement base that can be used when designing a CI monitoring system or a user interface to support SA. The requirements can also be used during system or user interface evaluation, and as a guide for what aspects to emphasize when training new CI monitoring operators. To create the SA requirements, goal-directed task analysis (GDTA) was conducted. For GDTA, nine interview sessions were held during the research. For a clear understanding of a CI monitoring operators work, all interviewees were subject matter experts (SMEs) and had extensive experience in CI monitoring. Before the interviews, a day-long observation session was conducted to gather initial input for the GDTA goal hierarchy and the SA requirements. GDTA identified three goals an operator must achieve in order to be successful in their work, and they were used to define the final SA requirements. As a result, a hierarchy diagram was constructed that includes three goals: monitoring, analysis and internal communication, and external communication. The SA requirements for a CI monitoring operator include information regarding ongoing incidents in the environment and the state of systems and services in the operators organization.

Modelling and real-time analysis of critical infrastructure using discrete event systems on graphs

Critical infrastructure (CI) systems form an interdependent network where failures in one system may quickly affect the state of other linked systems. Real-time modelling and analysis of CI systems gives valuable time-critical insight on the situational status during incidents and standard operation. Obtaining real-time quantitative measurements about the state of CI systems is necessary for situational awareness (SA) purposes. In this paper we present a general framework for real-time critical infrastructure modelling and analysis using discrete event systems (DES) on graphs. Our model augments standard graph-theoretic analysis with elements from automata theory to achieve model which captures interdependencies in CI. The framework was tested on various graphs with differing sizes and degree distributions. The resulting framework was implemented, and benchmarks indicate that it is suitable for real-time SA analysis.

Word and Phrase Dictionaries Generated with Multiple Translation Paths

Methods used to learn bilingual word embedding mappings, which project the source-language embeddings into the target embedding space, are compared in this paper. Orthogonal transformations, which are robust to noise, can learn to translate between word pairs they have never seen during training (zero-shot translation). Using multiple translation paths, e.g. Finnish \(\rightarrow \) English \(\rightarrow \) Russian and Finnish \(\rightarrow \) French \(\rightarrow \) Russian, at the same time and combining the results was found to improve the results of this process. Four new methods are presented for the calculation of either the single most similar or the five most similar words, based on the results of multiple translation paths. Of these, the Summation method was found to improve the P@1 translation precision by 1.6% points compared to the best result obtained with a direct translation (Fi \(\rightarrow \) Ru). The probability margin is presented as a confidence score. With similar coverages, the probability margin was found to outperform probability as a confidence score in terms of P@1 and P@5.

Text Augmentation Techniques for Document Vector Generation from Russian News Articles

In this paper, a document classification system is enhanced through the construction of a text augmentation technique by testing various Part-of-Speech filters and word vector weighting methods with nine different models for document representation. Subject/object tagging is introduced as a new form of text augmentation, along with a novel classification system grounded in a word weighting method based on the distribution of words among classes of documents. When an augmentation including subject/object tagging, a nouns+adjectives filter and Inverse Document Frequency word weighting was applied, an average increase in classification accuracy of 4.1% points was observed.

Integrated platform for critical infrastructure analysis and common operating picture solutions

In this paper, we present a software framework for modeling, simulation, and analysis of critical infrastructure (CI). Our concept fuses together a state-of-the-art telecommunications and electricity distribution system simulator (CI simulator), and a Common Operating Picture visualization system (COP system). The development process included expert interviews, which were conducted to define a comprehensive set of end-user requirements from different critical infrastructure stakeholders benefitting from a common situational picture. Using the obtained results, we enhanced the CI simulator to model more precisely interdependencies in communication and electricity distribution networks in normal and abnormal situations. In addition, the simulator was extended with near future prediction capabilities using the current situation and networks’ operating conditions. The simulator also provides a real-time data stream to the COP system, whose core analysis and visualization functions were specified according to the end-user requirements collected from the interviews.

Tactical Network Modeller simulation tool

This paper discusses the implementation of a tactical network simulation tool. The tool is called Tactical Network Modeller (TNM). TNM uses some novel techniques to simplify the building of the network model using graph theory constrained by a hierarchical tree which reflects the organisation structure. TNM allows models to be constructed using an Application Programming Interface (API) or a node based User Interface (UI). When the model is constructed, different simulation back-ends can be applied to it. A discrete event simulation and a network emulation back-end are implemented building on top of open source tools. TNM is simple to create models for non technical users. The model can be used to analyse information flows. The same model can be used for a full network emulation. This allows real software and protocols to be tested in a realistic simulated environment. The flexibility of the software allows its use from engineering up to campaign planning.