Assessing short answers in educational settings is challenging due to the need for scalability and accuracy, which led to the field of Automatic Short Answer Grading (ASAG). Traditional machine learning models, such as ensemble and embeddings, have been widely researched in ASAG, but they often suffer from generalizability issues. Recently, Large Language Models (LLMs) emerged as an alternative to optimize ASAG systems. However, previous research has failed to present a comprehensive analysis of LLMs' performance powered by prompt engineering strategies and compare its capabilities to traditional models. This study presents a comparative analysis between traditional machine learning models and GPT-4 in the context of ASAG. We investigated the effectiveness of different models and text representation techniques and explored prompt engineering strategies for LLMs. The results indicate that traditional machine learning models outperform LLMs. However, GPT-4 showed promising capabilities, especially when configured with optimized prompt components, such as few-shot examples and clear instructions. This study contributes to the literature by providing a detailed evaluation of LLM performance compared to traditional machine learning models in a multilingual ASAG context, offering insights for developing more efficient automatic grading systems.

Education for innovation initiatives often adopt Project-Based Learning approaches, where students are grouped into teams with the goal of creating an innovative product prototype. Working on teams, especially interdisciplinary ones, involves challenges related to collaboration and motivation [1]. It is well known that gamification is a successful tool for engaging students in learning activities [2]. Thus, in this paper, we propose a cardgame focused on education for innovation, exploring the similarities between the students’ teams with the roles of an RPG (Role Playing Game). To design our game proposal, we conducted a netnographic [3] and semi-structured qualitative research [4]. After implementing the game, we evaluated it in “Projetão,” a well-consolidated course on education for innovation. The results were encouraging.

Games are employing economic incentives, such as daily rewards, fidelity points, virtual goods and currencies, aiming to improve ARM funnel metrics related to Users Acquisition, Retention and Monetization. Massive Multiplayer Online Game (MMOG) go a step further creating virtual economies that increase the sense of ownership in players, improving retention [3]. Unfortunately, analyzing and managing virtual economies may be challenging, since there is a great amount and diversity of data to be considered. Economic indicators, largely used in the real economy, may be useful tools for analyzing MMOG economy, as well as conducting interventions. However, unveiling the economic indicators that are relevant ARM metrics is hard, since models for virtual economies presented in the literature do not address this issue. We identified only one research work proposing economic indicators for MMOG economies, but they are only two (inflation and nominal wages) and they have not been validated in any realtime game operation. In this paper, we propose six novel economic indicators for MMOG operators. We have preliminarily validated these indicators with experts; following we formalized the variables they depend on; then, we implemented a solution via a visual dashboard for a commercial MMOG. We collected the MMOG's data for the period of one year (six months before and six months after the use of the new indicators), covering 416,000 different players, with an average of 8,400 players per day. The results were amazingly positive. We obtained significant improvements in the main commercial metrics. Beyond its practical and solid results, we believe that another important contribution of our work is to draw the attention of developers into the interplay between economic science and games operation. We claim that the increasing introduction of economic elements in games should come together with the incorporation of concepts and tools for real world economics.

Algorithmic interventions (the intricate, dynamic arrangements of people and code deployed to address everyday life problems) benefit society. However, they are also responsible for many complex social, economic, and political harms. Unfortunately, there is an epistemological gap in computer science. Computer science lacks the language, the knowledge, and the methods for dealing with how to make algorithms’ stakeholders responsible, accountable, and liable for their outcomes’ effects, impacts, and consequences. Our research focuses on seizing pragmatism philosophy and linguistic pragmatics to reform computer science epistemology, expanding SDLC (system development life cycle framework) based on clinical pragmatics. We support reform to SDLC to help diminish algorithmic interventions’ adverse effects by design or, at least, to assign stakeholders’ duties for every algorithmic intervention in society. The algorithmic pragmatism conceptual framework presented could be taught in schools and universities and be used to public regulate and judge algorithmic interventions. Overall, we aim to design means of thinking about computer systems and algorithms to fill the computer science epistemology gap related to the impacts and consequences of algorithmic interventions.

The rapid adoption of Machine Learning (ML) in human’s daily lives and activities is raising ethical dilemmas and issues. A form of minimizing possible harm to society is to provide guidance to ML developers, who can build systems that are ethical by design. Unfortunately, developers do not have proper ethical formation in regular undergraduate courses, and the existing documents, despite being abundant, are vague and focused on governments and corporations rather than on individual developers. This paper proposes ethical recommendations, 18 concrete guidelines and 24 best practices, for developers. These recommendations were formulated in a focus group and validated quantitatively in a survey with over 130 ML developers working in both industry and Academia. This paper also investigates the state of adoption of such recommendations and compares what developers think they should do to achieve more ethical results versus what they actually do.

This paper presents four correlated studies focused on (i) finding out whether performance artists were interested in exploring interactive poetics based on mocap technologies, (ii) what the entry barriers were preventing such exploration and (iii) how to lower such barriers. Data from an original survey of one hundred and twelve (112) artists show that most of the respondents want to use motion capture (mocap) technologies to explore new poetic possibilities (Study 1). Only a few have had a chance to do so, however, because using the current mocap operating software often requires programming knowledge, and these programs do not offer built-in integration with mocap sensors (Study 2). This leads artists to depend on collaboration with IT experts, increasing the efforts and costs of their creative endeavors. We then executed a participatory design of 300 h with four dancers and one musician from a local contemporary dance/performance group (Study 3). This study resulted in a performance and some guidelines for lowering the usability entry barrier of such mocap-enabled systems, with a special focus on innovating experimentation with interactive poetics in the performing arts. Aiming to validate these guidelines, we built marine, an entry-level system that demonstrates one of the possible ways of implementing them. Following this design-prototyping-testing phase, a formal evaluation occurred based on comparative tests and interviews with ten dancers (Study 4). The results showed that marine improved effectiveness, efficiency, engagement and ease of learning, as well as provided a more pleasant user experience when compared to visual programming software. Data from an original survey conducted with 112 artists indicate that artists are interested in experimenting with motion capture technology, but usability problems represent a barrier to appropriation of these technologies for use in the performing arts. A participatory design, including a performance, led to unpublished guidelines for developing entry-level systems dedicated to helping artists explore new poetic possibilities with these technologies. A novel entry-level system called marine implements these guidelines, providing a user-friendly interface for artists.

A great number of techniques were already applied to the non-adversarial variation of the multiagent patrolling problem. Experiments suggest that, for general graphs, all those approaches are inferior to a strategy based on the travelling salesman problem (TSP), in which agents are distributed equidistantly along the TSP-cycle. This approach, however, is neither optimal nor scalable. In this article, we present a novel patrolling strategy which tries to overcome these limitations. Inspired on Newton's law of gravitation, our approach consists in assigning to each node of the graph an abstract mass which grows while the node remains unvisited, creating a force that attracts agents to it. We defined some variations of our approach and experimentally compared them to the TSP-based strategy, concluding that our approach performs better in general.

The development of new market-driven software products involves several challenges for the requirements engineering process. The challenges are deeper in the case these products are mobile video games. In particular, the mobile games must satisfy a number of critical non-functional requirements (e.g. portability, gameplay, emotional issues). In addition to that, mobile games are developed for mass market, demanding from the development team to understand the requirements of very diverse and sometimes unknown groups of stakeholders. This paper presents the experience of meantime, a global mobile game developing company, in conducting requirements engineering activities during the development of the Frogman game. It also presents some lessons learned in the process.

Porting is a critical task in mobile device game development. The high diversity of devices requires correspondingly customized versions of a single game. Managing the variabilities among these customized versions while exploring the latent game commonality cannot be solely addressed with a single technique. This paper contributes to this task by identifying and analyzing porting challenges, and by evaluating and contrasting existing approaches within industrial-strength case studies of J2ME games. Further, we present lessons learned, proposing more effective guidelines for this process, aiming at improving the quality of the resulting applications and porting process productivity.

Functional harmonic analysis is an important task in music composition, accompaniment, arrangement and others. However, the solutions are still not satisfactory. The proposed process is divided into two levels: the former extends one of previous works in the domain to carry out a richer analysis of chord grids and is where the very analysis is performed, and the latter is devoted to correct some conceptual inconsistencies concerning enharmonic spelling of chords. Both levels use an engine to make inferences on some rule bases, which can be easily improved by addition of new rules.

Ant colony optimization (ACO) is a metaheuristic inspired by the biological behavior of Argentine ants, which cooperate with each other by means of pheromone deposit and evaporation. The ant system (AS), one of the instances of this general optimization approach, has been applied to different optimization problems. However, the feasibility of the AS has not been experimentally evaluated on an important category of the facility location problem, called the set covering problem (SCP). The unique application of ACO to SCP reported by literature (Hadji et al., 2000) used ant colonies combined with a local search. However its evaluation was not designed to provide more general conclusive results. For this reason, this paper not only worked with a non-hybrid ACO algorithm, but also adopted a more accurate experimental evaluation method, where a descriptive analysis comes before a comparative evaluation.