Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

The power of feedback.

This paper reviews peer-reviewed empirical studies on gamification. We create a framework for examining the effects of gamification by drawing from the definitions of gamification and the discussion on motivational affordances. The literature review covers results, independent variables (examined motivational affordances), dependent variables (examined psychological/behavioral outcomes from gamification), the contexts of gamification, and types of studies performed on the gamified systems. The paper examines the state of current research on the topic and points out gaps in existing literature. The review indicates that gamification provides positive effects, however, the effects are greatly dependent on the context in which the gamification is being implemented, as well as on the users using it. The findings of the review provide insight for further studies as well as for the design of gamified systems.

/pdf/does-gamification-work-a-literature-review-of-empirical-3d9hukqa8z.pdf

Does Gamification Work? -- A Literature Review of Empirical Studies on Gamification

An overview of the self-organizing map algorithm, on which the papers in this issue are based, is presented in this article.

The Self-Organizing Map

Recommender systems have developed in parallel with the web. They were initially based on demographic, content-based and collaborative filtering. Currently, these systems are incorporating social information. In the future, they will use implicit, local and personal information from the Internet of things. This article provides an overview of recommender systems as well as collaborative filtering methods and algorithms; it also explains their evolution, provides an original classification for these systems, identifies areas of future implementation and develops certain areas selected for past, present or future importance.

http://romisatriawahono.net/lecture/rm/survey/information%20retrieval/Bobadilla%20-%20Recommender%20Systems%20-%202013.pdf

Recommender systems survey

Adaptive visualization approaches attempt to tune the content 
and the topology of information visualization to various user characteristics. While adapting visualization to user cognitive traits, goals, or 
knowledge has been relatively well explored, some other user characteristics have received no attention. This paper presents a methodology 
to adapt a traditional cluster-based visualization of communities to user 
individual model of community organization. This class of user-adapted 
visualization is not only achievable, but expected due to real world situation where users cannot be segmented into heterogeneous communities since many users have affinity to more than one group. An interactive clustering and visualization approach presented in the paper allows the user communicate their personal mental models of overlapping 
communities to the clustering algorithm itself and obtain a community 
visualization image that more realistically fits their prospects.

/pdf/adaptive-visualization-of-research-communities-i3y4bt34rf.pdf

Adaptive visualization of research communities

The emergence of generative large language models (LLMs) raises the question: what will be its impact on crowdsourcing. Traditionally, crowdsourcing has been used for acquiring solutions to a wide variety of human-intelligence tasks, including ones involving text generation, manipulation or evaluation. For some of these tasks, models like ChatGPT can potentially substitute human workers. In this study, we investigate, whether this is the case for the task of paraphrase generation for intent classification. We quasi-replicated the data collection methodology of an existing crowdsourcing study (similar scale, prompts and seed data) using ChatGPT. We show that ChatGPT-created paraphrases are more diverse and lead to more robust models.

ChatGPT to Replace Crowdsourcing of Paraphrases for Intent Classification: Higher Diversity and Comparable Model Robustness

. A powerful set of educational tools has emerged over the last decade with the rise in the adoption of online adaptive learning content. An increasingly popular tool in this space is the “intelligent textbook” as a platform to support and distribute content for e-learning, given its re-semblance with real-life physical books. Existing efforts in this direction include the development of digital textbooks where both textual content and interactive learning activities (i.e., examples, problems, etc.) are carefully handcrafted by the authors so that they are perfectly placed to follow the knowledge acquisition-practice flow. However, this approach is very time-consuming, and it requires the work of high-expertise authors. In this work, we suggest and discuss a scalable solution: we take existing digital textbooks and augment them by using repositories of existing online learning material associated with the subject matter. We present our current work in this direction and discuss challenges and opportunities for the future work.

Augmenting Digital Textbooks with Reusable Smart Learning Content: Solutions and Challenges

The constant increase in the amount of data and information available on the Web has made the development of systems that can support users in making relevant decisions increasingly important. Recommender systems (RSs) have emerged as tools to address this task. RSs use the preferences expressed by a user, either explicitly or implicitly, to filter the available information and proactively suggest items that might be of interest to him or her. Although in early works about the topic there was a strong interest in ways to make such systems proactive, user-friendly, and persuasive, over time they became increasingly focused on the algorithmic component solely. However, this trend is gradually being reversed and always more attention is nowadays placed also on Human Decision Making models that focus on supporting the end user in understanding what is being proposed through RSs by using dynamic and persuasive interfaces. A recommender system should be based on valuable strategies for proactively guiding users to items that match their preferences and therefore should put attention on how it is possible to make this process trustable, pleasant, and user-friendly. Such systems, moreover, should take into account psychological, cognitive and emotional aspects to enable personalization that is appropriate not only to the context of use but also to the psychological reactions of the end user. The workshop provides a venue for works that invest in the design of recommender systems which consider users’ experience during the interaction, as well as for works that explore the implications of human-computer interactions with different theories of human decision-making. In this summary, we introduce the Joint Workshop on Interfaces and Human Decision Making for Recommender Systems at RecSys’22, review its history, and discuss the most important topics considered at the workshop.

Joint Workshop on Interfaces and Human Decision Making for Recommender Systems (IntRS’22)

Personalized learning and educational recommender systems are integral parts of modern online education systems. In this context, the problem of recommending the best learning material to students is a perfect example of sequential multi-objective recommendation. Learning material recommenders need to optimize for and balance between multiple goals, such as adapting to student ability, adjusting the learning material difficulty, increasing student knowledge, and serving student interest, at every step of the student learning sequence. However, the obscurity and incompatibility of these objectives pose additional challenges for learning material recommenders. To address these challenges, we propose Proximity-based Educational Recommendation (PEAR), a recommendation framework that suggests a ranked list of problems by approximating and balancing between problem difficulty and student ability. To achieve an accurate approximation of these objectives, PEAR can integrate with any state-of-the-art student and domain knowledge model. As an example of such student and domain knowledge model, we introduce Deep Q-matrix based Knowledge Tracing model (DQKT), and integrate PEAR with it. Rather than static recommendations, this framework dynamically suggests new problems at each step by tracking student knowledge level over time. We use an offline evaluation framework, Robust Evaluation Matrix (REM), to compare PEAR with various baseline recommendation policies under three different student simulators and demonstrate the effectiveness of our proposed model. We experiment with different student trajectory lengths and show that while PEAR can perform better than the baseline policies with fewer data, it is also robust with longer sequence lengths. Student Knowledge Modeling. Both DQKT and DKVMN update a value matrix, M 𝑡𝑣 of size C × 𝑑 ℎ , as the model observes student attempts on problems. This matrix can be interpreted as the student’s mastery levels of each concept feature, at time step 𝑡 . In DQKT, to represent student knowledge in each concept as a C -dimensional vector k 𝑡𝑢 ∈ R C , we summarize the value matrix M 𝑡𝑣 by summing over the concept latent dimension. The value matrix M 𝑡𝑣 could be updated with an erase-follow-by-add operation as shown in the write component, which is exactly the same as in the DKVMN.

Peter Brusilovsky

Papers

Adaptive visualization of research communities

ChatGPT to Replace Crowdsourcing of Paraphrases for Intent Classification: Higher Diversity and Comparable Model Robustness

Augmenting Digital Textbooks with Reusable Smart Learning Content: Solutions and Challenges

Joint Workshop on Interfaces and Human Decision Making for Recommender Systems (IntRS’22)

Proximity-Based Educational Recommendations: A Multi-Objective Framework