Automatically grading the diverse range of question types in high school physics problem is a challenge that requires automated grading techniques from different fields. We report the findings of a Systematic Literature Review of potential physics grading techniques. We propose a multi-modal AI grading framework to address these challenges and examine our framework in light of Australia's AI Ethical Principles.
Reinhard Moratz, Niklas Daute, James Ondieki
et al.
This paper deals with improving the capabilities of Large Language Models (LLM) to provide route instructions for pedestrian wayfinders by means of qualitative spatial relations.
The advent of conversational agents with increasingly human-like behaviour throws old philosophical questions into new light. Does it, or could it, ever make sense to speak of AI agents built out of generative language models in terms of consciousness, given that they are "mere" simulacra of human behaviour, and that what they do can be seen as "merely" role play? Drawing on the later writings of Wittgenstein, this paper attempts to tackle this question while avoiding the pitfalls of dualistic thinking.
The article further develops and formalizes a theory of friendly dialogue in an AI System of Dr. Watson type, as proposed in our previous publication[4],[19]. The main principle of this type of AI is to guide the user toward a solution in a friendly manner, using questions based on the analysis of user input and data collected in the system.
Case studies (CS) attempt to help students increase critical thinking skills and engagement while working through a real-life scenario in various disciplines, including medicine, law, and business. However, the CS method has not been heavily utilized in biological sciences. The present study investigated the effect of the CS method on undergraduate biology students’ conceptual understanding, academic outcomes, and perspectives. A case study was applied in a one-semester undergraduate biology course, which was compared to ten semesters of standard sections. Participants completed course pre- and post-tests, pre- and post-case tests, and an online survey to assess their conceptual understanding and engagement. The initial lowest quartiles were determined from the individual course pre-test scores, which were lower than class averages. Results suggested that the CS method helped students in learning outcomes, critical thinking, and conceptual understanding toward biology. In post-test learning gains, the CS group did 20% better than the non-CS group, with the largest benefit seen in the initially lowest pre-test quartile of the class. Moreover, post-case learning gains were 55% improved in the case test. Survey results indicated that students had positive attitudes toward CS for their engagement in plant biology content. Overall, the distribution of A grades improved by 2.6-fold from standard to CS groups. We conclude that the use of CS may address course content engagement and have the potential to effectively boost academic performance, especially for the initially lowest quartile in undergraduate plant biological sciences courses.
Marc Pierre, Quentin Cohen-Solal, Tristan Cazenave
Monte Carlo Tree Search can be used for automated theorem proving. Holophrasm is a neural theorem prover using MCTS combined with neural networks for the policy and the evaluation. In this paper we propose to improve the performance of the Holophrasm theorem prover using other game tree search algorithms.
The Newcomb's paradox is one of the most known paradox in Game Theory about the Oracles. We will define the graph associated to the time lines of the Game. After this Studying its topology and using only the Expected Utility Principle we will formulate a solution of the paradox able to explain all the classical cases.
Roland Fernandez, Asli Celikyilmaz, Rishabh Singh
et al.
We present a formal language with expressions denoting general symbol structures and queries which access information in those structures. A sequence-to-sequence network processing this language learns to encode symbol structures and query them. The learned representation (approximately) shares a simple linearity property with theoretical techniques for performing this task.
We define and study a general framework for approval-based budgeting methods and compare certain methods within this framework by their axiomatic and computational properties. Furthermore, we visualize their behavior on certain Euclidean distributions and analyze them experimentally.
I propose a system for Automated Theorem Proving in higher order logic using deep learning and eschewing hand-constructed features. Holophrasm exploits the formalism of the Metamath language and explores partial proof trees using a neural-network-augmented bandit algorithm and a sequence-to-sequence model for action enumeration. The system proves 14% of its test theorems from Metamath's set.mm module.
The elegant Stalnaker/Lewis semantics for counterfactual conditonals works with distances between models. But human beings certainly have no tables of models and distances in their head. We begin here an investigation using a more realistic picture, based on findings in neuroscience. We call it a pre-semantics, as its meaning is not a description of the world, but of the brain, whose structure is (partly) determined by the world it reasons about. In the final section, we reconsider the components, and postulate that there are no atomic pictures, we can always look inside.
Relational representations in reinforcement learning allow for the use of structural information like the presence of objects and relationships between them in the description of value functions. Through this paper, we show that such representations allow for the inclusion of background knowledge that qualitatively describes a state and can be used to design agents that demonstrate learning behavior in domains with large state and actions spaces such as computer games.
Despite the prevalence of the Computational Theory of Mind and the Connectionist Model, the establishing of the key principles of the Cognitive Science are still controversy and inconclusive. This paper proposes the concept of Pattern Recognition as Necessary and Sufficient Principle for a general cognitive science modeling, in a very ambitious scientific proposal. A formal physical definition of the pattern recognition concept is also proposed to solve many key conceptual gaps on the field.
We address the problem of propositional logic-based abduction, i.e., the problem of searching for a best explanation for a given propositional observation according to a given propositional knowledge base. We give a general algorithm, based on the notion of projection; then we study restrictions over the representations of the knowledge base and of the query, and find new polynomial classes of abduction problems.
We have an audacious dream, we would like to develop a simulation and virtual reality system to support the decision making in European football (soccer). In this review, we summarize the efforts that we have made to fulfil this dream until recently. In addition, an introductory version of FerSML (Footballer and Football Simulation Markup Language) is presented in this paper.
For natural and artificial systems with some symmetry structure, computational understanding and manipulation can be achieved without learning by exploiting the algebraic structure. Here we describe this algebraic coordinatization method and apply it to permutation puzzles. Coordinatization yields a structural understanding, not just solutions for the puzzles.
We present a domain-independent algorithm that computes macros in a novel way. Our algorithm computes macros "on-the-fly" for a given set of states and does not require previously learned or inferred information, nor prior domain knowledge. The algorithm is used to define new domain-independent tractable classes of classical planning that are proved to include \emph{Blocksworld-arm} and \emph{Towers of Hanoi}.
We consider the well-known family ALC(D) of description logics with a concrete domain, and provide first results on a framework obtained by augmenting ALC(D) atemporal roles and aspatial concrete domain with temporal roles and a spatial concrete domain.
We define a quantitative constraint language subsuming two calculi well-known in QSR (Qualitative Spatial Reasoning): Frank's cone-shaped and projection-based calculi of cardinal direction relations. We show how to solve a CSP (Constraint Satisfaction Problem) expressed in the language.
Information personalization is fertile ground for application of AI techniques. In this article I relate personalization to the ability to capture partial information in an information-seeking interaction. The specific focus is on personalizing interactions at web sites. Using ideas from partial evaluation and explanation-based generalization, I present a modeling methodology for reasoning about personalization. This approach helps identify seven tiers of `personable traits' in web sites.