Cognitive Evaluation of Machine Learning Agents

Advances in applying statistical Machine Learning (ML) led to several claims of human-level or near-human performance in tasks such as image classification & speech recognition. Such claims are unscientific primarily for two reasons, (1) They incorrectly enforce the notion that task-specific performance can be treated as manifestation of General Intelligence and (2) They are not verifiable as currently there is no set benchmark for measuring human-like cognition in a machine learning agent. Moreover, ML agent’s performance is influenced by knowledge ingested in it by its human designers. Therefore, agent’s performance may not necessarily reflect its true cognition. In this paper, we propose a framework that draws parallels from human cognition to measure machine’s cognition. Human cognitive learning is quite well studied in developmental psychology with frameworks and metrics in place to measure actual learning. To either believe or refute the claims of human-level performance of machine learning agent, we need scientific methodology to measure its cognition. Our framework formalizes incremental implementation of human-like cognitive processes in ML agents with an implicit goal to measure it. The framework offers guiding principles for measuring, (1) Task-specific machine cognition and (2) General machine cognition that spans across tasks. The framework also provides guidelines for building domain-specific task taxonomies to cognitively profile tasks. We demonstrate application of the framework with a case study where two ML agents that perform Vision and NLP tasks are cognitively evaluated.     

Does change in ethical education influence core moral values? Towards history- and culture-aware morality model with application in automatic moral reasoning

In this study, we focus on ethical education as a means to improve artificial companion’s conceptualization of moral decision-making process in human users. In particular, we focus on automatically determining whether changes in ethical education influenced core moral values in humans throughout the century. We analyze ethics as taught in Japan before WWII and today to verify how much the pre-WWII moral attitudes have in common with those of contemporary Japanese, to what degree what is taught as ethics in school overlaps with the general population’s understanding of ethics, as well as to verify whether a major reform of the guidelines for teaching the school subject of “ethics” at school after 1946 has changed the way common people approach core moral questions (such as those concerning the sacredness of human life). We selected textbooks used in teaching ethics at school from between 1935 and 1937, and those used in junior high schools today (2019) and analyzed what emotional and moral associations such contents generated. The analysis was performed with an automatic moral and emotional reasoning agent and based on the largest available text corpus in Japanese as well as on the resources of a Japanese digital library. As a result, we found out that, despite changes in stereotypical view on Japan’s moral sentiments, especially due to historical events, past and contemporary Japanese share a similar moral evaluation of certain basic moral concepts, although there is a large discrepancy between how they perceive some actions to be beneficial to the society as a whole while at the same time being inconclusive when it comes to assessing the same action’s outcome on the individual performing them and in terms of emotional consequences. Some ethical categories, assessed positively before the war, while being associated with a nationalistic trend in education have also disappeared from the scope of interest of post- war society. The findings of this study support suggestions proposed by others that the development of personal AI systems requires supplementation with moral reasoning. Moreover, the paper builds upon this idea and further suggests that AI systems need to be aware of ethics not as a constant, but as a function with a correction on historical and cultural changes in moral reasoning.     

心理因素与学业表现：机器学习分类预测模型

随着高等教育规模的扩大,学业表现不良逐渐成为一个不容忽视的现象,对学业表现不良的学生进行预测并提早给予干预可降低退学率并减少教育资源的损失。由于导致学业表现不良的因素众多且关系复杂,传统的基于相关分析的研究方法很难建立早期预测模型并进行应用。本研究旨在利用机器学习算法,对数据进行挖掘,并建立学业表现预测模型。研究对653名大一新生的心理健康状况、应对方式、人格、内外控倾向和相关人口统计学信息进行了收集,并在一年后采集了其学业成绩,利用随机森林(RF)、K邻近（KNN）、支持向量机（SVM）、决策树（DT）、朴素贝叶斯（NB）等机器学习算法建立了分类模型。结果显示,随机森林算法在识别学业表现不良学生时有最好的表现,其中准确率95.86%, 召回率91.83%,f1分数为93.80%。特征重要性分析显示,前10个对模型有最高贡献度的特征包括：年龄、性别、是否为独生子、内外控倾向、神经质倾向、积极应对倾向、宜人性倾向、一般症状指数、开放性倾向和焦虑水平。为避免过度拟合问题,本研究在一年后收集的166名新生样本中进行了模型验证,结果显示模型在新数据集上有较好的泛化表现,其中f1分数90.90％,准确率92.60％,召回率89.26％。研究提示基于人口统计学和心理测评信息,机器学习算法有助于及早识别学业表现不良学生并为开展早期干预提供启示。     

Toward ethical cognitive architectures for the development of artificial moral agents

New technologies based on artificial agents promise to change the next generation of autonomous systems and therefore our interaction with them. Systems based on artificial agents such as self-driving cars and social robots are examples of this technology that is seeking to improve the quality of people’s life. Cognitive architectures aim to create some of the most challenging artificial agents commonly known as bio-inspired cognitive agents. This type of artificial agent seeks to embody human-like intelligence in order to operate and solve problems in the real world as humans do. Moreover, some cognitive architectures such as Soar, LIDA, ACT-R, and iCub try to be fundamental architectures for the Artificial General Intelligence model of human cognition. Therefore, researchers in the machine ethics field face ethical questions related to what mechanisms an artificial agent must have for making moral decisions in order to ensure that their actions are always ethically right. This paper aims to identify some challenges that researchers need to solve in order to create ethical cognitive architectures. These cognitive architectures are characterized by the capacity to endow artificial agents with appropriate mechanisms to exhibit explicit ethical behavior. Additionally, we offer some reasons to develop ethical cognitive architectures. We hope that this study can be useful to guide future research on ethical cognitive architectures.     

On Modal Logics Characterized by Models with Relative Accessibility Relations: Part I

This work is divided in two papers (Part I and Part II). In Part I, we study a class of polymodal logics (herein called the class of "Rare-logics") for which the set of terms indexing the modal operators are hierarchized in two levels: the set of Boolean terms and the set of terms built upon the set of Boolean terms. By investigating different algebraic properties satisfied by the models of the Rare-logics, reductions for decidability are established by faithfully translating the Rare-logics into more standard modal logics. The main idea of the translation consists in eliminating the Boolean terms by taking advantage of the components construction and in using various properties of the classes of semilattices involved in the semantics. The novelty of our approach allows us to prove new decidability results (presented in Part II), in particular for information logics derived from rough set theory and we open new perspectives to define proof systems for such logics (presented also in Part II).     

The prospects for a Jungian psychology in China

The possibility of a Jungian psychology developing in China is considered by a brief historical excursion through the early translations of psychoanalytical works. Translation problems and the contentious nature of some of Freud's ideas have made for their difficult reception in China. The inattention to Jung's ideas in universities in the west in the past, and a reliance on science based subjects by Chinese students studying abroad, have meant little opportunity to study Jung, and, by implication, to translate him. The turbulent political climate in China over the course of the past century has also hindered developments in psychology generally. In addition, certain traditional practices of understanding mind-body relationships and reporting 'illnesses' have precluded the possibility of any psychotherapeutic psychology emerging. However, the present climate looks more favourable for the dissemination of Jungian knowledge, but the question of an appropriate context and a receptive readership remains. Certain Jungian notions can be seen to fit comfortably within traditional systems of Chinese thought but the present day psychology department in China is no more a congenial environment for Jungian psychology than its counterpart in the west. It may be that the success of importing Jungian ideas into China rests with those with a predilection towards arts and cultural sciences, and with the innovations of the organizers of conferences.     

Sociotechnical Imaginaries and the Globalization of Converging Technology Policy: Technological Developmentalism in South Korea

The United States (US), European Union (EU), and South Korea had different definitions and visions of technological convergence before interacting with each other from the late 2000s. The Korean government has used Western policies as a benchmark but produces a distinct concept of technological convergence due to a particular imaginary of technology as a vehicle for national economic growth. This sociotechnical imaginary of technological developmentalism influences Korea’s translation of technological convergence from other jurisdictions. Because the sociotechnical imaginaries of different nations are difficult to communicate across national contexts, the translation of visions often changes the original meaning of things like technological convergence. Western sociotechnical imaginaries such as human enhancement and sustainable development do not translate well into South Korea due to the national imaginary of technological developmentalism. First, Korea’s Lee administration predominantly envisioned converging technology (CT) as a new growth engine, in contrast to the US and EU which emphasize visions of trans-humanist or sustainable futures respectively. Second, the Park administration’s CT vision imitates the Western societal challenge-driven rationale, but this vision was not enacted. As such, the democratic, sustainable imaginary of societal challenge-driven innovation is not easily translated into Korea’s national imaginary or technology policy.     

An fMRI study of the social competition in healthy subjects

Social interaction requires the ability to infer another person’s mental state (Theory of Mind, ToM) and also executive functions. This fMRI study aimed to identify the cerebral correlates activated by ToM during a specific social interaction, the human-human competition. In this framework, we tested a conflict resolution task (Stroop) adapted to a virtual situation of competition. The participants were instructed to play in order to win either against a human-like competitor (human-human competition) or against a non-human competitor (human-machine competition). Only the human-human competition requires ToM as this type of competition is performed under social interaction. We identified first the classical network of executive regions activated by Stroop. Secondly, we identified the social (human-human) competition regions, represented by the bilateral superior and inferior frontal gyri, the anterior cingulate, the insula, the superior and anterior temporal, the hippocampus, the fusiform gyrus, the cuneus and the precuneus. Finally, we identified the executive regions that were modulated by the human-human competition, i.e., the executive control regions additionally activated when mentalizing in the context of social competition. They constituted a network predominant to the right and composed of the superior and middle frontal, anterior cingulate, insula and fusiform gyrus. We suggest that our experimental paradigm may be useful in exploration of the cerebral correlates of social adjustments in several situations such as psychiatric disorders presenting executive and social dysfunctions.     

Mind,Machine, and Creativity: An Artist's Perspective

Harold Cohen is a renowned painter who has developed a computer program, AARON, to create art. While AARON has been hailed as one of the most creative AI programs, Cohen consistently rejects the claims of machine creativity. Questioning the possibility for AI to model human creativity, Cohen suggests in so many words that the human mind takes a different route to creativity, a route that privileges the relational, rather than the computational, dimension of cognition. This unique perspective on the tangled web of mind, machine, and creativity is explored by an application of three relational models of the mind to an analysis of Cohen's talks and writings, which are available on his website: www.aaronshome.com .     

Italian Adaptation and Confirmatory Factor Analysis of the Full and the Short Form of the Posttraumatic Growth Inventory

This study was designed to adapt and to assess the reliability and factor structure of an Italian translation of the full and short form of the Posttraumatic Growth Inventory. A sample of 1,244 Italian adults who had experienced a range of adverse life events participated in this study. Five models of the underlying structure of the Posttraumatic Growth Inventory were tested via confirmatory factor analyses. Analyses of both the short and full form of the PTGI provided support for the original correlated five-factor structure. Multigroup confirmatory factor analysis supported the invariance of the Posttraumatic Growth Inventory across gender.