Search

Gli algoritmi sono vulnerabili a sviluppare pregiudizi che potrebbero rendere le loro decisioni ingiuste nei confronti di particolari gruppi di individui. La fairness comporta una serie di aspetti che dipendono fortemente dal dominio di applicazione che considera diverse nozioni su cosa sia una decisione giusta nelle situazioni che influiscono sugli individui della popolazione. Le precise differenze, implicazioni e ortogonalità tra queste nozioni non sono state ancora completamente analizzate e il lavoro prova a mattere ordini in questo zoo di definizioni. Quando si tratta di far rispettare tali vincoli, la maggior parte dei modelli di mitigazione in-processing incorpora vincoli di fairness come componente fondamentale della loss function e richiede quindi aggiustamenti a livello di codice per adattarsi a contesti specifici e domini. Piuttosto che fare affidamento su un approccio procedurale, il nostro modello sfrutta la struttura dichiarativa conoscenza acquisita per codificare i requisiti di equità sotto forma di regole logiche che catturano affermazioni del linguaggio naturale bigue e precise. Proponiamo un'integrazione neuro-simbolica basato su Logic Tensor Networks che combina apprendimento di reti nueurali basato sui daticonoscenze logiche di alto livello, consentendo di eseguire task di classificazione riducendo discriminazione. Le prove sperimentali mostrano che le prestazioni sono buone quanto quelle dello stato dell’arte fornendo così un quadro flessibile per tenere conto della non discriminazione, spesso a un costo modesto in termini di precisione., Algorithms are vulnerable to biases that might render their decisions unfair toward particular groups of individuals. Fairness comes with a range of facets that strongly depend on the application domain that consider different notions of what is a fair decision in situations impacting individuals in the population. The precise differences, implications and orthogonality between these notions have not yet been fully analyzed and we try to make some order out of this zoo of definitions. When it comes about enforcing such constraints, most in-processing mitigation models embed fairness constraints as fundamental component of the loss function thus requiring code-level adjustments to adapt to specific contexts and domains. Rather than relying on a procedural approach, our model leverages declarative structured knowledge to encode fairness requirements in the form of logic rules capturing unam- biguous and precise natural language statements. We propose a neuro-symbolic integration approach based on Logic Tensor Networks that combines data-driven network-based learning with high-level logical knowledge, allowing to perform classification tasks while reducing discrimination. Experimental evidence shows that performance is as good as state-of-the-art thus providing a flexible framework to account for non-discrimination often at a modest cost in terms of accuracy.

Data availability is becoming more essential, considering the current growth of web-based data. The data available on the web are represented as unstructured, semi-structured, or structured data. In order to make the web-based data available for several Natural Language Processing or Data Mining tasks, the data needs to be presented as machine-readable data in a structured format. Thus, techniques for addressing the problem of capturing knowledge from unstructured data sources are needed. Knowledge extraction methods are used by the research communities to address this problem; methods that are able to capture knowledge in a natural language text and map the extracted knowledge to existing knowledge presented in knowledge graphs (KGs). These knowledge extraction methods include Named-entity recognition, Named-entity Disambiguation, Relation Recognition, and Relation Linking. This thesis addresses the problem of extracting knowledge over unstructured data and discovering patterns in the extracted knowledge. We devise a rule-based approach for entity and relation recognition and linking. The defined approach effectively maps entities and relations within a text to their resources in a target KG. Additionally, it overcomes the challenges of recognizing and linking entities and relations to a specific KG by employing devised catalogs of linguistic and domain-specific rules that state the criteria to recognize entities in a sentence of a particular language, and a deductive database that encodes knowledge in community-maintained KGs. Moreover, we define a Neuro-symbolic approach for the tasks of knowledge extraction in encyclopedic and domain-specific domains; it combines symbolic and sub-symbolic components to overcome the challenges of entity recognition and linking and the limitation of the availability of training data while maintaining the accuracy of recognizing and linking entities. Additionally, we present a context-aware framework for unveiling semantically rela

We introduce CLEVR-Math, a multi-modal math word problems dataset consisting of simple math word problems involving addition/subtraction, represented partly by a textual description and partly by an image illustrating the scenario. The text describes actions performed on the scene that is depicted in the image. Since the question posed may not be about the scene in the image, but about the state of the scene before or after the actions are applied, the solver envision or imagine the state changes due to these actions. Solving these word problems requires a combination of language, visual and mathematical reasoning. We apply state-of-the-art neural and neuro-symbolic models for visual question answering on CLEVR-Math and empirically evaluate their performances. Our results show how neither method generalise to chains of operations. We discuss the limitations of the two in addressing the task of multi-modal word problem solving.

We introduce CLEVR-Math, a multi-modal math word problems dataset consisting of simple math word problems involving addition/subtraction, represented partly by a textual description and partly by an image illustrating the scenario. The text describes actions performed on the scene that is depicted in the image. Since the question posed may not be about the scene in the image, but about the state of the scene before or after the actions are applied, the solver envision or imagine the state changes due to these actions. Solving these word problems requires a combination of language, visual and mathematical reasoning. We apply state-of-the-art neural and neuro-symbolic models for visual question answering on CLEVR-Math and empirically evaluate their performances. Our results show how neither method generalise to chains of operations. We discuss the limitations of the two in addressing the task of multi-modal word problem solving.

We introduce CLEVR-Math, a multi-modal math word problems dataset consisting of simple math word problems involving addition/subtraction, represented partly by a textual description and partly by an image illustrating the scenario. The text describes actions performed on the scene that is depicted in the image. Since the question posed may not be about the scene in the image, but about the state of the scene before or after the actions are applied, the solver envision or imagine the state changes due to these actions. Solving these word problems requires a combination of language, visual and mathematical reasoning. We apply state-of-the-art neural and neuro-symbolic models for visual question answering on CLEVR-Math and empirically evaluate their performances. Our results show how neither method generalise to chains of operations. We discuss the limitations of the two in addressing the task of multi-modal word problem solving.

We introduce CLEVR-Math, a multi-modal math word problems dataset consisting of simple math word problems involving addition/subtraction, represented partly by a textual description and partly by an image illustrating the scenario. The text describes actions performed on the scene that is depicted in the image. Since the question posed may not be about the scene in the image, but about the state of the scene before or after the actions are applied, the solver envision or imagine the state changes due to these actions. Solving these word problems requires a combination of language, visual and mathematical reasoning. We apply state-of-the-art neural and neuro-symbolic models for visual question answering on CLEVR-Math and empirically evaluate their performances. Our results show how neither method generalise to chains of operations. We discuss the limitations of the two in addressing the task of multi-modal word problem solving.

We introduce CLEVR-Math, a multi-modal math word problems dataset consisting of simple math word problems involving addition/subtraction, represented partly by a textual description and partly by an image illustrating the scenario. The text describes actions performed on the scene that is depicted in the image. Since the question posed may not be about the scene in the image, but about the state of the scene before or after the actions are applied, the solver envision or imagine the state changes due to these actions. Solving these word problems requires a combination of language, visual and mathematical reasoning. We apply state-of-the-art neural and neuro-symbolic models for visual question answering on CLEVR-Math and empirically evaluate their performances. Our results show how neither method generalise to chains of operations. We discuss the limitations of the two in addressing the task of multi-modal word problem solving.

We introduce CLEVR-Math, a multi-modal math word problems dataset consisting of simple math word problems involving addition/subtraction, represented partly by a textual description and partly by an image illustrating the scenario. The text describes actions performed on the scene that is depicted in the image. Since the question posed may not be about the scene in the image, but about the state of the scene before or after the actions are applied, the solver envision or imagine the state changes due to these actions. Solving these word problems requires a combination of language, visual and mathematical reasoning. We apply state-of-the-art neural and neuro-symbolic models for visual question answering on CLEVR-Math and empirically evaluate their performances. Our results show how neither method generalise to chains of operations. We discuss the limitations of the two in addressing the task of multi-modal word problem solving.

We introduce CLEVR-Math, a multi-modal math word problems dataset consisting of simple math word problems involving addition/subtraction, represented partly by a textual description and partly by an image illustrating the scenario. The text describes actions performed on the scene that is depicted in the image. Since the question posed may not be about the scene in the image, but about the state of the scene before or after the actions are applied, the solver envision or imagine the state changes due to these actions. Solving these word problems requires a combination of language, visual and mathematical reasoning. We apply state-of-the-art neural and neuro-symbolic models for visual question answering on CLEVR-Math and empirically evaluate their performances. Our results show how neither method generalise to chains of operations. We discuss the limitations of the two in addressing the task of multi-modal word problem solving.

We introduce CLEVR-Math, a multi-modal math word problems dataset consisting of simple math word problems involving addition/subtraction, represented partly by a textual description and partly by an image illustrating the scenario. The text describes actions performed on the scene that is depicted in the image. Since the question posed may not be about the scene in the image, but about the state of the scene before or after the actions are applied, the solver envision or imagine the state changes due to these actions. Solving these word problems requires a combination of language, visual and mathematical reasoning. We apply state-of-the-art neural and neuro-symbolic models for visual question answering on CLEVR-Math and empirically evaluate their performances. Our results show how neither method generalise to chains of operations. We discuss the limitations of the two in addressing the task of multi-modal word problem solving.

We introduce CLEVR-Math, a multi-modal math word problems dataset consisting of simple math word problems involving addition/subtraction, represented partly by a textual description and partly by an image illustrating the scenario. The text describes actions performed on the scene that is depicted in the image. Since the question posed may not be about the scene in the image, but about the state of the scene before or after the actions are applied, the solver envision or imagine the state changes due to these actions. Solving these word problems requires a combination of language, visual and mathematical reasoning. We apply state-of-the-art neural and neuro-symbolic models for visual question answering on CLEVR-Math and empirically evaluate their performances. Our results show how neither method generalise to chains of operations. We discuss the limitations of the two in addressing the task of multi-modal word problem solving.

We introduce CLEVR-Math, a multi-modal math word problems dataset consisting of simple math word problems involving addition/subtraction, represented partly by a textual description and partly by an image illustrating the scenario. The text describes actions performed on the scene that is depicted in the image. Since the question posed may not be about the scene in the image, but about the state of the scene before or after the actions are applied, the solver envision or imagine the state changes due to these actions. Solving these word problems requires a combination of language, visual and mathematical reasoning. We apply state-of-the-art neural and neuro-symbolic models for visual question answering on CLEVR-Math and empirically evaluate their performances. Our results show how neither method generalise to chains of operations. We discuss the limitations of the two in addressing the task of multi-modal word problem solving.

Logic program and neural networks are two important aspects in artificial intelligence. This paper is part of an endeavour towards neural networks and logic programming integration. The goal in performing logic programming based on the energy minimization scheme is to achieve the best ratio of global minimum. However, there is no guarantee to find the best minimum in the network. To achieve this, a new learning rule based Activation Function was derived to accelerate the performance of doing logic programming in Hopfield Neural Network (HNN). This paper also shows focused on agent based modelling for presenting performance of doing logic programming in Hopfield network using new activation function. The effects of the activation function are analyzed mathematically and compared with the existing method. Computer simulations are carried out by using NETLOGO to validate the effectiveness on the new activation function. The resuls obtained showed that the new activation function outperform the existing method in doing logic programming in Hopfield network. The models developed by agent based modelling also support this theory.

Logic program and neural networks are two important aspects in artificial intelligence. This paper is part of an endeavour towards neural networks and logic programming integration. The goal in performing logic programming based on the energy minimization scheme is to achieve the best ratio of global minimum. However, there is no guarantee to find the best minimum in the network. To achieve this, a new learning rule based Activation Function was derived to accelerate the performance of doing logic programming in Hopfield Neural Network (HNN). This paper also shows focused on agent based modelling for presenting performance of doing logic programming in Hopfield network using new activation function. The effects of the activation function are analyzed mathematically and compared with the existing method. Computer simulations are carried out by using NETLOGO to validate the effectiveness on the new activation function. The resuls obtained showed that the new activation function outperform the existing method in doing logic programming in Hopfield network. The models developed by agent based modelling also support this theory.