Generating Predictable and Adaptive Dialog Policies in Single- and Multi-domain Goal-oriented Dialog Systems

Most existing commercial goal-oriented chatbots are diagram-based; i.e. they follow a rigid dialog flow to fill the slot values needed to achieve a user’s goal. Diagram-based chatbots are predictable, thus their adoption in commercial settings; however, their lack of flexibility may cause many users to leave the conversation before achieving their goal. On the other hand, state-of-the-art research chatbots use Reinforcement Learning (RL) to generate flexible dialog policies. However, such chatbots can be unpredictable, may violate the intended business constraints, and require large training datasets to produce a mature policy. We propose a framework that achieves a middle ground between the diagram-based and RL-based chatbots: we constrain the space of possible chatbot responses using a novel structure, the chatbot dependency graph, and use RL to dynamically select the best valid responses. Dependency graphs are directed graphs that conveniently express a chatbot’s logic by defining the dependencies among slots: all valid dialog flows are encapsulated in one dependency graph. Our experiments in both single-domain and multi-domain settings show that our framework quickly adapts to user characteristics and achieves up to 23.77% improved success rate compared to a state-of-the-art RL model.

A graph-based framework for malicious software detection and classification utilizing temporal-graphs

In this paper we present a graph-based framework that, utilizing relations between groups of System-calls, detects whether an unknown software sample is malicious or benign, and classifies a malicious software to one of a set of known malware families. In our approach we propose a novel graph representation of dependency graphs by capturing their structural evolution over time constructing sequential graph instances, the so-called Temporal Graphs. The partitions of the temporal evolution of a graph defined by specific time-slots, results to different types of graphs representations based upon the information we capture across the capturing of its evolution. The proposed graph-based framework utilizes the proposed types of temporal graphs computing similarity metrics over various graph characteristics in order to conduct the malware detection and classification procedures. Finally, we evaluate the detection rates and the classification ability of our proposed graph-based framework conducting a series of experiments over a set of known malware samples pre-classified into malware families.

Analysing Mission-critical Cyber-physical Systems with AND/OR Graphs and MaxSAT

Cyber-Physical Systems (CPS) often involve complex networks of interconnected software and hardware components that are logically combined to achieve a common goal or mission; for example, keeping a plane in the air or providing energy to a city. Failures in these components may jeopardise the mission of the system. Therefore, identifying the minimal set of critical CPS components that is most likely to fail, and prevent the global system from accomplishing its mission, becomes essential to ensure reliability. In this article, we present a novel approach to identifying the Most Likely Mission-critical Component Set (MLMCS) using AND/OR dependency graphs enriched with independent failure probabilities. We address the MLMCS problem as a Maximum Satisfiability (MaxSAT) problem. We translate probabilities into a negative logarithmic space to linearise the problem within MaxSAT. The experimental results conducted with our open source tool LDA4CPS indicate that the approach is both effective and efficient. We also present a case study on complex aircraft systems that shows the feasibility of our approach and its applicability to mission-critical cyber-physical systems. Finally, we present two MLMCS-based security applications focused on system hardening and forensic investigations.

Application of the Dependency Graph Method in the Analysis of Automatic Transmission Gearboxes

The current article discusses the use of the dependency graph method to design and analyze automatic transmissions. Different goals may be served by modeling an automatic transmission using graphs. The most important of them are: determining the gear ratios for gears and the analysis of speed and acceleration of the rotating elements. Game tree-structure methods can be used to analyze the functional schemes of the selected gears. This paper describes the method of generating a system of kinematics equations for signal dependency graphs. This allows the generalization and extension of the algorithmic approach, and also further analyses and syntheses, such as checking the isomorphism of the proposed solutions and determining the validity of the structure and the operational parameters of the analyzed gears.

Change in the Complexity of the Studied Concepts and Theoretical Models with the Age of the Student

The problem of the dependence of the didactic complexity of the studied concepts and theoretical models on the age of the student (schoolchild, student) is analysed. The complexity of a concept (term) can be characterized by the number of words from a fifth-grader’s thesaurus needed to explain its meaning. To find the complexity of a theoretical model of an object (for example, an atom), it is necessary to sum up the complexities of all the words that make up the description of the model and take into account the indicator of the variety of terms. Dependency graphs were built: 1) the complexity of the most difficult terms for understanding from the year of study; 2) the complexity of various theoretical models of the atom from the year of study at school and university. In both cases, the resulting curves are ascending, like a parabola, corresponding to an increase in complexity by almost a hundred times.