Transformer models, trained and publicly released over the last couple of years, have proved effective in many NLP tasks. We wished to test their usefulness in particular on the stance detection task. We performed experiments on the data from the Fake News Challenge Stage 1 (FNC-1). We were indeed able to improve the reported SotA on the challenge, by exploiting the generalization power of large language models based on Transformer architecture. Specifically (1) we improved the FNC-1 best performing model adding BERT sentence embedding of input sequences as a model feature, (2) we fine-tuned BERT, XLNet, and RoBERTa transformers on FNC-1 extended dataset and obtained state-of-the-art results on FNC-1 task.

Network log events produced by network probes are used by security analyzers to detect traffic anomalies and threats. While it is relatively trivial for a probe to mark specific events as suspicious, it is much more challenging for log analyzers to create a comprehensive picture of the overall network. Machine learning can potentially help in this, however there is no specific solution for analyzing network event logs. This paper covers the experiments and design choices that have been made to create a machine learning-based tool able to analyze network event logs. The tool has been evaluated in the Suspicious Network Event Recognition Cup at IEEE BigData 2019, achieving an AUC (Area Under the Curve) of over 90%, making it accurate enough for deployment in real life scenarios.

In April 2018, under the auspices of the POR-FESR 2014-2020 program of Italian Piedmont Region, the Turin's Centre on High-Performance Computing for Artificial Intelligence (HPC4AI) was funded with a capital investment of 4.5M€ and it began its deployment. HPC4AI aims to facilitate scientific research and engineering in the areas of Artificial Intelligence and Big Data Analytics. HPC4AI will specifically focus on methods for the on-demand provisioning of AI and BDA Cloud services to the regional and national industrial community, which includes the large regional ecosystem of Small-Medium Enterprises (SMEs) active in many different sectors such as automotive, aerospace, mechatronics, manufacturing, health and agrifood.

Antonio Valerio Miceli-Barone, Giuseppe Attardi. Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2015.

We discuss the problem of ranking very many entities of different types. In particular we deal with a heterogeneous set of types, some being very generic and some very specific. We discuss two approaches for this problem: i) exploiting the entity containment graph and ii) using a Web search engine to compute entity relevance. We evaluate these approaches on the real task of ranking Wikipedia entities typed with a state-of-the-art named-entity tagger. Results show that both approaches can greatly increase the performance of methods based only on passage retrieval.

We present a framework for code generation that allows programs to manipulate and generate code at the source level while the joining and splicing of executable code is carried out automatically at the intermediate code/VM level. The framework introduces a data type Code to represent code fragments: methods/operators from this class are used to reify a method from a class, producing its representation as an object of type Code. Code objects can be combined by partial application to other Code objects. Code combinators, corresponding to higher-order methods, allow splicing the code of a functional actual parameter into the resulting Code object. CodeBricks is a library implementing the framework for the .NET Common Language Runtime. The framework can be exploited by language designers to implement metaprogramming, multistage programming and other language features. We illustrate the use of the technique in the implementation of a fully featured regular expression compiler that generates code emulating a finite state automaton. We present benchmarks comparing the performance of the RE matcher built with CodeBricks with the hand written one present in .NET.

Software c omponents encourage code reuse and simplify application development.An increasing number of applications is built assembling components developed by third parties, taking advantage of language-independence, object orientation, ease of use and other features of modern component architectures.Computer algebra systems could exploit the software component approach, but several issues must be addressed, mostly due to the sophisticated data structures required for representing mathematical objects.We discuss these problems and present a proposal based on the OpenMath specications.We built an prototype f r amework for developing and using mathematical components.The framework uses IDL from CORBA for specifying the interfaces for objects.Code developed in the framework is mapped into either the COM object model for creating ActiveX components or into CORBA objects for creating servers implemented as dynamic modules.

Automatic garbage collection relieves programmers from the burden of managing memory themselves and several techniques have been developed that make garbage collection feasible in many situations, including real time applications or within traditional programming languages. However, optimal performance cannot always be achieved by a uniform general purpose solution. Sometimes an algorithm exhibits a predictable pattern of memory usage that could be better handled specifically, delaying as much as possible the intervention of the general purpose collector. This leads to the requirement for algorithm specific customisation of the collector strategies. We present a dynamic memory management framework which can be customised to the needs of an algorithm, while preserving the convenience of automatic collection in the normal case. The Customisable Memory Manager (CMM) organises memory in multiple heaps. Each heap is an instance of C++ class which abstracts and encapsulates a particular storage discipline. The default heap for collectable objects uses the technique of mostly copying garbage collection, providing good performance and memory compaction. Customisation of the collector is achieved exploiting object orientation by defining specialised versions of the collector methods for each heap class. The object-oriented interface to the collector enables coexistence and coordination among the various collectors as well as integration with traditional code unaware of garbage collection. The CMM is implemented in C++ without any special support in the language or the compiler. The techniques used in the CMM are general enough to be applicable also to other languages. The performance of the CMM is analysed and compared to other conservative collectors for C/C++ in various configurations. © 1998 John Wiley & Sons, Ltd.

We present a formalisation for the notion of viewpoint, a construct meant for expressing several varieties of relativised truth. The formalisation consists in a logic which extends first order predicate calculus with its own metalanguage, an axiomatization of provability and proper reflection rules. The extension is not conservative, but consistency is granted. Viewpoints are defined as set of reified meta-level sentences. A proof theory for viewponts is developed which enables to carry out proofs of statements involving several viewpoints. A semantic account of viewpoints is provided, dealing with issues of self referential theories and paradoxes, and exploiting the notion of contextual entailment. Notions such as beliefs, knowledge, absolute truth or truth relative to a situation can be uniformly modeled as provability in specialised viewpoints, obtained by imposing suitable constraints on viewpoints.

We discuss the advantages of using a logic system for knowledge representation which is based on descriptions, rather than predicates, and which embodies two fundamental ideas for structuring knowledge that are distilled from semantic networks and frame-based languages: inheritance and attributions. Taxonomic reasoning on a lattice of descriptions combined with deduction strategies defined at the metalevel provide the knowledge base with the capability to deal with complex problem-solving tasks.