Education students engage in diverse learning activities requiring appropriate assistance and timely feedback. As their numbers grow, providing them with scalable support is an important challenge. Here, we focus on t...
ISBN:
(纸本)9783031723117;9783031723124
Education students engage in diverse learning activities requiring appropriate assistance and timely feedback. As their numbers grow, providing them with scalable support is an important challenge. Here, we focus on the development of a didactic chatbot based on a Large Language Model (LLM). The potential of LLMs is enhanced by existing materials and pedagogical course descriptions. Using Retrieval Augmented Generation (RAG), the bot can retrieve and analyse course materials, in order to provide comprehensive answers to specific questions. Preliminary results indicate that it is possible to distinguish between different student contexts and to generate a prompt answer, taking into account the relevant materials. The evaluation results achieved 84.78% accuracy in providing correct answers for seminar materials.
Learning is a social process. However, online learning is characterized by social isolation. We present an AI social actor, SAMI, for fostering social interactions in online classrooms. SAMI (Social Agent Mediated Int...
ISBN:
(纸本)9783031630279;9783031630286
Learning is a social process. However, online learning is characterized by social isolation. We present an AI social actor, SAMI, for fostering social interactions in online classrooms. SAMI (Social Agent Mediated Interactions) aims to mitigate the potentially adverse impact of social isolation on the learning experience and emotional well-being of geographically dispersed online learners in asynchronous educational settings. SAMI connects learners based on their shared identity considering student location, hobbies, and academic interests. It also aims to enhance the feeling of "belongingness" felt by the students in the community of online students. SAMI has been deployed at Georgia Institute of Technology in several online classes with over 11000 students in the past two years. We describe our findings from student surveys to gauge SAMI's effectiveness.
Labeling style affects labeling efficiency and quality in image annotation tasks. For example, a "label quickly" style can increase labeling efficiency when the data are easy, and a "label carefully&quo...
ISBN:
(纸本)9783031601064;9783031601071
Labeling style affects labeling efficiency and quality in image annotation tasks. For example, a "label quickly" style can increase labeling efficiency when the data are easy, and a "label carefully" style can increase label quality when the data are difficult. However, the selection of an appropriate labeling style is difficult as different annotators have different experiences and domain knowledge, affecting their subjective feelings of data difficulties (for example, User 1 feels Data A to be easy, while User 2 feels it difficult). In this paper, we propose "Dynamic Labeling" as a control system for labeling styles used in image-labeling tasks. Our control system analyzes the labeling behaviors of annotators (i.e., label selection time) and dynamically assigns an appropriate labeling style (label quickly or label carefully). We conducted a user study to compare a conventional "nondynamic" and the proposed "dynamic" labeling approaches for an image-labeling task. The results suggest that Dynamic Labeling increased the label quality and labeling efficiency.
We design efficient and robust algorithms for the batch posting of rollup chain calldata on the base layer chain, using tools from operations research. We relate the costs of posting and delaying, by converting them t...
ISBN:
(纸本)9783031488054;9783031488061
We design efficient and robust algorithms for the batch posting of rollup chain calldata on the base layer chain, using tools from operations research. We relate the costs of posting and delaying, by converting them to the same units and adding them up. The algorithm that keeps the average and maximum queued number of batches tolerable enough improves the posting costs of the trivial algorithm, which posts batches immediately when they are created, by 8%. On the other hand, the algorithm that only cares moderately about the batch queue length can improve the trivial algorithm posting costs by 29%. Our findings can be used by layer two projects that post data to the base layer at some regular rate.
The development of accurate causal models is crucial for achieving and explaining desired outcomes that require interventions. Building these models efficiently requires combining available data with expert causal kno...
ISBN:
(纸本)9783031722400;9783031722417
The development of accurate causal models is crucial for achieving and explaining desired outcomes that require interventions. Building these models efficiently requires combining available data with expert causal knowledge. Often experts have unique data and model insights, but sharing them is challenging due to privacy or security concerns. Federated machine learning addresses similar issues by allowing multiple sites to collaborate on a common model without sharing private datasets. This paper introduces CCaT, a distributed causal discovery tool enabling collaborative development of a shared causal model while preserving local models and data privacy. CCaT allows each site to evaluate and refine the shared model using its private dataset, sharing only summary statistics or suggested new causal relations. The tool supports maintaining distinct local causal models, as analysts can choose to adopt or change parts of the shared model. CCaT enhances the accuracy of causal models by leveraging diverse expertise and data, achieving a generality and accuracy unattainable by individual sites. We present several common scenarios with the CCaT to demonstrate its effectiveness.
Recently, the computational complexity of the Pumping-Problem, that is, for a given finite automaton A and a value p, deciding whether the language L(A) satisfies a previously fixed regular pumping lemma w.r.t. the va...
ISBN:
(数字)9783031661594
ISBN:
(纸本)9783031661587;9783031661594
Recently, the computational complexity of the Pumping-Problem, that is, for a given finite automaton A and a value p, deciding whether the language L(A) satisfies a previously fixed regular pumping lemma w.r.t. the value p, was considered in [H. Gruber and M. Holzer and C. Rauch. The Pumping Lemma for Regular Languages is Hard. CIAA 2023, pp. 128-140.]. Here we generalize the Pumping-Problem by investigating Bar-Hillel's context-free pumping lemma instead. It turns out that for context-free languages, the Pumping-Problem for Bar-Hillel's pumping lemma is undecidable. When restricted to regular languages, the problem under consideration becomes decidable.
Pyramid schemes are investment scams in which top-level participants in a hierarchical network recruit and profit from an expanding base of defrauded newer participants. They have existed for over a century, but their...
ISBN:
(纸本)9783031477508;9783031477515
Pyramid schemes are investment scams in which top-level participants in a hierarchical network recruit and profit from an expanding base of defrauded newer participants. They have existed for over a century, but their historical opacity has prevented in-depth studies. This paper presents an empirical study of Forsage, a smart-contract-based pyramid scheme with unprecedented transparency. Our study focuses on the period around 2020, when Forsage was one of the largest contracts (by gas usage) in Ethereum. In 2022, some months after initial release of this work, the U.S. SEC dubbed Forsage a "fraudulent crypto pyramid and Ponzi scheme" and filed charges against its creators and promoters. We quantify the (multi-million-dollar) gains of top-level participants as well as the losses of the vast majority (around 88%) of users. We analyze Forsage code both manually and using a purpose-built transaction simulator that we release as open source software to uncover the complex mechanics of the scheme. Through complementary study of promotional videos and social media, we show how Forsage promoters leveraged the unique features of smart contracts to lure users with false claims of trust-worthiness and profitability, and how Forsage activity is concentrated within a small number of national communities. Our analysis is the most complete study of a pyramid scheme to date.
In this paper, we introduce a family of games called concave pro-rata games. In such a game, players place their assets into a pool, and the pool pays out some concave function of all assets placed into it. Each playe...
ISBN:
(纸本)9783031488054;9783031488061
In this paper, we introduce a family of games called concave pro-rata games. In such a game, players place their assets into a pool, and the pool pays out some concave function of all assets placed into it. Each player then receives a pro-rata share of the payout;i.e., each player receives an amount proportional to how much they placed in the pool. Such games appear in a number of practical scenarios, including as a simplified version of batched decentralized exchanges, such as those proposed by Penumbra. We show that this game has a number of interesting properties, including a symmetric pure equilibrium that is the unique equilibrium of this game, and we prove that its price of anarchy is O(n) in the number of players. We also show some numerical results in the iterated setting which suggest that players quickly converge to an equilibrium in iterated play.
Elections where electors rank the candidates (or a subset of the candidates) in order of preference allow the collection of more information about the electors' intent. The most widely used election of this type i...
ISBN:
(纸本)9783031488054;9783031488061
Elections where electors rank the candidates (or a subset of the candidates) in order of preference allow the collection of more information about the electors' intent. The most widely used election of this type is Instant-Runoff Voting (IRV), where candidates are eliminated one by one, until a single candidate holds the majority of the remaining ballots. Condorcet elections treat the election as a set of simultaneous decisions about each pair of candidates. The Condorcet winner is the candidate who beats all others in these pairwise contests. There are various proposals to determine a winner if no Condorcet winner exists. In this paper we show how we can efficiently audit Condorcet elections for a number of variations. We also compare the audit efficiency (how many ballots we expect to sample) of IRV and Condorcet elections.
This study aims to support Project-Based Learning (PBL) in a distance environment in which teachers provide feedback to learners to facilitate learning. However, in distance PBL, it is difficult for teachers to grasp ...
ISBN:
(纸本)9783031679971;9783031679988
This study aims to support Project-Based Learning (PBL) in a distance environment in which teachers provide feedback to learners to facilitate learning. However, in distance PBL, it is difficult for teachers to grasp learners' learning status, and it is difficult for them to provide appropriate feedback. In this study, to assist teachers in providing feedback to learners whose learning status is not good, we estimate engagement, which is one of the mental indicators of their positive attitude toward learning, using activity reports submitted by the learners. In this paper, we analyzed the content of activity reports from perspectives other than the occurrence of negative words, which has been reported on previously. The results of the analysis suggest the possibility of estimating learner engagement and learning status through sentiment estimation using a widely used sentiment analysis program, indicating that activity reports are useful in assisting teachers in providing feedback.
暂无评论