In this paper we introduce, ROCCA for Rational OpenCog Controlled Agent, an agent, that, as its name suggests, leverages the OpenCog framework to fulfill goals in uncertain environments. It attempts to act rationally,...
ISBN:
(数字)9783031334696
ISBN:
(纸本)9783031334689;9783031334696
In this paper we introduce, ROCCA for Rational OpenCog Controlled Agent, an agent, that, as its name suggests, leverages the OpenCog framework to fulfill goals in uncertain environments. It attempts to act rationally, relying on reasoning for both learning and planning. An experiment in a Minecraft environment is provided as a test case.
In the nuclear industry, the need for improved reliability in current and future technology hinders the deployment of autonomous robotic systems. The following research aims to develop a method of reliably mapping a l...
ISBN:
(数字)9783031433603
ISBN:
(纸本)9783031433597;9783031433603
In the nuclear industry, the need for improved reliability in current and future technology hinders the deployment of autonomous robotic systems. The following research aims to develop a method of reliably mapping a large environment and abstracting the map into a sparse node graph to create a more efficient data form. The proposed data form allows for efficient storage whilst maintaining important map features and coverage. The method utilises an expanding node algorithm to convert standard occupancy maps to a sparse node graph representation. The algorithm's effectiveness has been tested on simulated maps and real-world maps to test the compression factor for a wide range of scenarios. The algorithm is expanded to function on a semi-unknown map abstracting during exploration.
Data-driven methods, machine learning and artificialintelligence methods are not yet exploited to their intended potential in solving the technical-technological challenges, especially in industrial applications, des...
详细信息
ISBN:
(纸本)9783031358906;9783031358913
Data-driven methods, machine learning and artificialintelligence methods are not yet exploited to their intended potential in solving the technical-technological challenges, especially in industrial applications, despite versatile development progress. This is mainly justified by the insufficient practicality of AI solutions. To exploit the potential of AI methods, technical practitioners often rely on interdisciplinary collaboration with data science specialists or consultants. In any development and application of AI methods, a plethora of methods must be mastered for solution-oriented acquisition, pre-processing and quality assurance of required data, as well as for the selection of suitable algorithms and their adaptation. Coping with this complexity usually requires a great deal of effort, both for the individual domain expert and for the data engineers and data analysts. Complexity and intransparency of AI methods therefore hinder the effectiveness and efficiency of AI deployment. Focusing on user-friendly delivery of AI-based applications, the paradigm of Usable AI (UAI) has been defined. This paper first summarizes the UAI paradigm. Finally, some application examples from the field of production engineering illustrate how UAI can improve the practical applicability of AI methods for domain experts.
The rise of Alzheimer's Disease worldwide has prompted a search for efficient tools which can be used to predict deterioration in cognitive decline leading to dementia. In this paper, we explore the potential of s...
ISBN:
(纸本)9783031414558;9783031414565
The rise of Alzheimer's Disease worldwide has prompted a search for efficient tools which can be used to predict deterioration in cognitive decline leading to dementia. In this paper, we explore the potential of survival machine learning as such a tool for building models capable of predicting not only deterioration but also the likely time to deterioration. We demonstrate good predictive ability (0.86 C-Index), lending support to its use in clinical investigation and prediction of Alzheimer's Disease risk.
The Stable diffusion [1] is a system composed of three parts - text encoder, latent diffusion model and autoencoder decoder. With the open source of Stable diffusion, more and more users begin to use stable diffusion ...
ISBN:
(纸本)9783031358906;9783031358913
The Stable diffusion [1] is a system composed of three parts - text encoder, latent diffusion model and autoencoder decoder. With the open source of Stable diffusion, more and more users begin to use stable diffusion to generate digital art, modify images and explore more applications. However, the application potential of stable diffusion in the different stages of industrial design is not yet clear. We divide the process of industrial design into four stages and focus on exploring its application in the sketching stage and rendering stage. We discussed whether the Stable-diffusion model can well express the concepts related to industrial design (product category, shape, color, material), explored the composability of the different finetune ways, and enabled Stable diffusion model to effectively transform the text prompt and image prompt into high-quality design scheme. It shows that finetuned stable diffusion model can help designers to build intent map and push structure deduction work. Also, with simple image hints and text prompt, finetuned stable diffusion model which trained from a specific product can do attribute, background and illumination modifications to the renderings.
Evolutionary multitasking optimization is a newly proposed paradigm that leverages the implicit parallelism of population-based search and transfers genetic information among tasks to solve multifactorial optimization...
ISBN:
(纸本)9783031402913;9783031402920
Evolutionary multitasking optimization is a newly proposed paradigm that leverages the implicit parallelism of population-based search and transfers genetic information among tasks to solve multifactorial optimization (MFO) problems simultaneously. However, simply migrating implicit genetics is incapable of fully exploring the potential of transfer learning to sum up crucial information and avoid local optima traps. Moreover, the negative transfer will hinder the search for the appropriate solution, leading to a performance collapse of the optimization algorithm. To address those issues, a multifactorial evolutionary algorithm based on model knowledge transfer (MT-MFEA) is proposed for MFO problems in this paper. In MT-MFEA, a clustering model is built based on density-based spatial clustering of applications with noise (DBSCAN) algorithm, which is beneficial to better capturing the position information of the search population for each task. Furthermore, explicit transfer learning is conducted among tasks by selecting higher-quality information from clustering models to enhance searchability and avoid negative transfer. The MT-MFEA is evaluated on various benchmark problems, involving optimization tasks with varying levels of intersecting global optima and similarities. The experimental analysis of the performances and comparisons with four state-of-the-art algorithms demonstrate the efficacy and efficiency of MT-MFEA.
Testing for Conditional Independence (CI) is a fundamental task for causal discovery but is particularly challenging in mixed discrete-continuous data. In this context, inadequate assumptions or discretization of cont...
ISBN:
(纸本)9783031434112;9783031434129
Testing for Conditional Independence (CI) is a fundamental task for causal discovery but is particularly challenging in mixed discrete-continuous data. In this context, inadequate assumptions or discretization of continuous variables reduce the CI test's statistical power, which yields incorrect learned causal structures. In this work, we present a non-parametric CI test leveraging k-nearest neighbor (kNN) methods that are adaptive to mixed discrete-continuous data. In particular, a kNN-based conditional mutual information estimator serves as the test statistic, and the p-value is calculated using a kNN-based local permutation scheme. We prove the CI test's statistical validity and power in mixed discrete-continuous data, which yields consistency when used in constraint-based causal discovery. An extensive evaluation of synthetic and real-world data shows that the proposed CI test outperforms state-of-the-art approaches in the accuracy of CI testing and causal discovery, particularly in settings with low sample sizes.
Adversarial defenses protect machine learning models from adversarial attacks, but are often tailored to one type of model or attack. The lack of information on unknown potential attacks makes detecting adversarial ex...
ISBN:
(纸本)9783031333736;9783031333743
Adversarial defenses protect machine learning models from adversarial attacks, but are often tailored to one type of model or attack. The lack of information on unknown potential attacks makes detecting adversarial examples challenging. Additionally, attackers do not need to follow the rules made by the defender. To address this problem, we take inspiration from the concept of Applicability Domain in cheminformatics. Cheminformatics models struggle to make accurate predictions because only a limited number of compounds are known and available for training. Applicability Domain defines a domain based on the known compounds and rejects any unknown compound that falls outside the domain. Similarly, adversarial examples start as harmless inputs, but can be manipulated to evade reliable classification by moving outside the domain of the classifier. We are the first to identify the similarity between Applicability Domain and adversarial detection. Instead of focusing on unknown attacks, we focus on what is known, the training data. We propose a simple yet robust triple-stage data-driven framework that checks the input globally and locally, and confirms that they are coherent with the model's output. This framework can be applied to any classification model and is not limited to specific attacks. We demonstrate these three stages work as one unit, effectively detecting various attacks, even for a white-box scenario.
Given a formal context, an ordinal factor is a subset of its incidence relation that forms a chain in the concept lattice, i.e., a part of the dataset that corresponds to a linear order. To visualize the data in a for...
ISBN:
(数字)9783031409608
ISBN:
(纸本)9783031409592;9783031409608
Given a formal context, an ordinal factor is a subset of its incidence relation that forms a chain in the concept lattice, i.e., a part of the dataset that corresponds to a linear order. To visualize the data in a formal context, Ganter and Glodeanu proposed a biplot based on two ordinal factors. For the biplot to be useful, it is important that these factors comprise as much data points as possible, i.e., that they cover a large part of the incidence relation. In this work, we investigate such ordinal two-factorizations. First, we investigate for formal contexts that omit ordinal two-factorizations the disjointness of the two factors. Then, we show that deciding on the existence of two-factorizations of a given size is an NP-complete problem which makes computing maximal factorizations computationally expensive. Finally, we provide the algorithm Ord2Factor that allows us to compute large ordinal two-factorizations.
Cooking soup is a complex dynamic process, where the properties and taste of ingredients change during long temperature exposure. Furthermore, the simmering process of a soup also causes evaporation of the water, whic...
ISBN:
(数字)9783031433603
ISBN:
(纸本)9783031433597;9783031433603
Cooking soup is a complex dynamic process, where the properties and taste of ingredients change during long temperature exposure. Furthermore, the simmering process of a soup also causes evaporation of the water, which increases the salt density in a bouillon. To mitigate this problem, we developed a closed-loop robotic system that allows cooking soups based on salinity and pH sensing. By taking into account that both salinity and pH are subject to change during the cooking, we recorded the salinity and pH over a complete course of cooking by an expert human and employed a proportional controller that adds salt and water into the soup. For the evaluation, we employed the proposed approach to cook a tomato soup with three different initial conditions. The results suggest that the system reaches the target pH and salinity reasonably close, even for significantly different soup bases.
暂无评论