Learning discriminative sequence models from partially labelled data for activity recognition

Recognising daily activity patterns of people from low-level sensory data is an important problem. Traditional approaches typically rely on generative models such as the hidden Markov models and training on fully labelled data. While activity data can be readily acquired from pervasive sensors, e.g. in smart environments, providing manual labels to support fully supervised learning is often expensive. In this paper, we propose a new approach based on partially-supervised training of discriminative sequence models such as the conditional random field (CRF) and the maximum entropy Markov model (MEMM). We show that the approach can reduce labelling effort, and at the same time, provides us with the flexibility and accuracy of the discriminative framework. Our experimental results in the video surveillance domain illustrate that these models can perform better than their generative counterpart (i.e. the partially hidden Markov model), even when a substantial amount of labels are unavailable.

Picturing evaporation : learning science literacy through a particle representation

Evaporation is mostly taught in primary schools through a water cycle representation. This has its limitations in explaining mechanisms and local effects such as drops drying in a closed room, condensation on cold surfaces, or how we smell liquids. In this paper the authors describe a classroom sequence of activities for Grade 5 students that explored the use of a particle model in conjunction with a range of representational modes, to explain evaporation phenomena. In interviews the authors explored with students their visual and verbal accounts of particles, modelling a process of teacher-mediated negotiation of multiple representations. From the evidence, the authors argue that difficulties in understanding evaporation are inherently representational, and that by engaging with the multiple literacies of science teachers can support significant advances in conceptual learning.

The application of a reinforcement learning agent to a multi-product manufacturing facility

An intelligent agent-based scheduling system, consisting of a reinforcement learning agent and a simulation model has been developed and tested on a classic scheduling problem. The production facility studied is a multiproduct serial line subject to stochastic failure. The agent goal is to minimise total production costs, through selection of job sequence and batch size. To explore state space the agent used reinforcement learning. By applying an independent inventory control policy for each product, the agent successfully identified optimal operating policies for a real production facility.

Learning from facial aging patterns for automatic age estimation

Age Specific Human-Computer Interaction (ASHCI) has vast potential applications in daily life. However, automatic age estimation technique is still underdeveloped. One of the main reasons is that the aging effects on human faces present several unique characteristics which make age estimation a challenging task that requires non-standard classification approaches. According to the speciality of the facial aging effects, this paper proposes the AGES (AGing pattErn Sub-space) method for automatic age estimation. The basic idea is to model the aging pattern, which is defined as a sequence of personal aging face images, by learning a representative subspace. The proper aging pattern for an unseen face image is then determined by the projection in the subspace that can best reconstruct the face image, while the position of the face image in that aging pattern will indicate its age. The AGES method has shown encouraging performance in the comparative experiments either as an age estimator or as an age range estimator.

Procedural learning in children with developmental coordination disorder

Despite the fact that developmental coordination disorder (DCD) is characterised by a deficit in the ability to learn or automate motor skills, few studies have examined motor learning over repeated trials. In this study we examined procedural learning in a group of 10 children with DCD (aged 8–12 years) and age-matched controls without DCD. The learning task was modelled on that of Nissen and Bullemer [Cognitive Psychology 19 (1987) 1]. Children performed a serial reaction time (SRT) task in which they were required to learn a spatial sequence that repeated itself every 10 trials. Children were not aware of the repetition. Spatial targets were four (horizontal) locations presented on a computer monitor. Children responded using four response keys with the same horizontal mapping as the stimulus. They were tested over five blocks of 100 trials each. The first four blocks presented the same repeating sequence, while the fifth block was randomised. Procedural learning was indexed by the slope of the regression of RT on blocks 1–4. Results showed that most children displayed strong procedural learning of the sequence, despite having no explicit knowledge about it. Overall, there was no group difference in the magnitude of learning over blocks of trials – most children performed within the normal range. Procedural learning for simple sequential movements appears to be intact in children with DCD. This suggests that cortico-striatal circuits that are strongly implicated in the sequencing of simple movements appear to be function normally in DCD.

Representational issues in students learning about evaporation

This study draws on recent research on the central role of representation in learning. While there has been considerable research on students’ understanding of evaporation, the representational issues entailed in this understanding have not been investigated in depth. The study explored students’ engagement with evaporation phenomena through various representational modes. The study indicates how a focus on representation can provide fresh insights into the conceptual task involved in learning science through an investigation of students’ responses to a structured classroom sequence and subsequent interviews over a year. A case study of one child’s learning demonstrates the way conceptual advances are integrally connected with the development of representational modes. The findings suggest that teacher-mediated negotiation of representational issues as students construct different modal accounts can support enriched learning by enabling both (a) richer conceptual understanding by students, and (b) enhanced teacher insights into students’ thinking.

Teaching and learning force as a representational issue : insights from a classroom video study

An enormous amount of research in the conceptual change tradition has shown the difficulty of learning fundamental science concepts, yet conceptual change schemes have failed to convincingly demonstrate improvements in supporting significant student learning. Recent work in cognitive science has challenged this purely conceptual view of learning, emphasising the role of languages, and the importance of personal and contextual aspects of  understanding science. The research described in this paper is designed around the notion that learning involves the recognition and development of students’ representational resources. In particular, we argue that difficulties with the concept of force are fundamentally representational in nature. The paper describes the planning and implementation of a classroom sequence in force that focuses on representations and their negotiation, and reports on the effectiveness of this perspective in guiding teaching and learning. Classroom sequences involving three teachers 158 2008 NARST Annual International Conference were videotaped using a combined focus on the teacher and groups of students. Video analysis software was used to code the variety of representations used, and sequences of representational negotiation. Stimulated recall interviews were conducted with teachers and students. The paper will report on the effect of this approach on teacher knowledge and pedagogy, and on student learning of force.

Multiple representation in learning about evaporation

There has been extensive research on children's understanding of evaporation, but representational issues entailed in this understanding have not been investigated in depth. This study explored three students' engagement with science concepts relating to evaporation through various representational modes, such as diagrams, verbal accounts, gestures, and captioned drawings. This engagement entailed students (a) clarifying their thinking through exploring representational resources; (b) developing understanding of what these representations signify; and (c) learning how to construct representational aspects of scientific explanation. The study involved a sequence of classroom lessons on evaporation and structured interviews with nine children, and found that a focus on representational challenges provided fresh insights into the conceptual task involved in learning science. The findings suggest that teacher-mediated negotiation of representational issues as students construct different modal accounts can support enriched learning by enabling both (a) richer conceptual understanding by students; and (b) enhanced teacher insights into students' thinking.

An explicit representational focus for teaching and learning about animals in the environment

There has been growing interest in linking the learning of Science with the literacies of Science and representations. Recent attention has been focused on learning theories that emphasise the socio-cultural and situated aspects of learning, and in particular the notion of learning as participation in a discourse community. This paper will describe a learning sequence planned wilh Year 5/6 teachers to study invertebrates in the schoolground environment, but with an additional focus in which students generated and negotiated representations, and discussed the adequacy of these. The paper will present data from video capture of classroom activities, students' work samples, and pre- and post-unit testing, to explore what a representational focus might entail in teaching science, and the role of representations in learning, reasoning and exploring in science.

Teaching and learning about force with a representational focus : pedagogy and teacher change

A large body of research in the conceptual change tradition has shown the difficulty of learning fundamental science concepts, yet conceptual change schemes have failed to convincingly demonstrate improvements in supporting significant student learning. Recent work in cognitive science has challenged this purely conceptual view of learning, emphasising the role of language, and the importance of personal and contextual aspects of understanding science. The research described in this paper is designed around the notion that learning involves the recognition and development of students’ representational resources. In particular, we argue that conceptual difficulties with the concept of force are fundamentally representational in nature. This paper describes a classroom sequence in force that focuses on representations and their negotiation, and reports on the effectiveness of this perspective in guiding teaching, and in providing insight into student learning. Classroom sequences involving three teachers were videotaped using a combined focus on the teacher and groups of students. Video analysis software was used to capture the variety of representations used, and sequences of representational negotiation. Stimulated recall interviews were conducted with teachers and students. The paper reports on the nature of the pedagogies developed as part of this representational focus, its effectiveness in supporting student learning, and on the pedagogical and epistemological challenges negotiated by teachers in implementing this approach.

Learning through constructing representations in science : a framework of representational construction affordances

Compared with research on the role of student engagement with expert representations in learning science, investigation of the use and theoretical justification of student-generated representations to learn science is less common. In this paper, we present a framework that aims to integrate three perspectives to explain how and why representational construction supports learning in science. The first or semiotic perspective focuses on student use of particular features of symbolic and material tools to make meanings in science. The second or epistemic perspective focuses on how this representational construction relates to the broader picture of knowledge-building practices of inquiry in this disciplinary field, and the third or epistemological perspective focuses on how and what students can know through engaging in the challenge of representing causal accounts through these semiotic tools. We argue that each perspective entails productive constraints on students’ meaning-making as they construct and interpret their own representations. Our framework seeks to take into account the interplay of diverse cultural and cognitive resources students use in these meaning-making processes. We outline the basis for this framework before illustrating its explanatory value through a sequence of lessons on the topic of evaporation.

Learning people movement model from multiple cameras for behaviour recognition

In surveillance systems for monitoring people behaviours, it is important to build systems that can adapt to the signatures of people's tasks and movements in the environment. At the same time, it is important to cope with noisy observations produced by a set of cameras with possibly different characteristics. In previous work, we have implemented a distributed surveillance system designed for complex indoor environments [1]. The system uses the Abstract Hidden Markov mEmory Model (AHMEM) for modelling and specifying complex human behaviours that can take place in the environment. Given a sequence of observations from a set of cameras, the system employs approximate probabilistic inference to compute the likelihood of different possible behaviours in real-time. This paper describes the techniques that can be used to learn the different camera noise models and the human movement models to be used in this system. The system is able to monitor and classify people behaviours as data is being gathered, and we provide classification results showing the system is able to identify behaviours of people from their movement signatures.

The role of representation in teaching and learning ideas about matter

The research described in this paper is designed around the notion that learning involves the recognition and development of students’ representational resources. This paper describes a classroom sequence in Ideas about Matter that focuses on representations and their negotiation, and reports on the effectiveness of this perspective in guiding teaching, and in providing further insight into student learning. Classroom sequences involving two experienced teachers (2008, Year 8 students) and an inexperienced teacher (2010, Year 7 students) were videotaped using a combined focus on the teacher and groups of students. Video analysis software was used to code the variety of representations used teachers and students, and sequences of representational negotiation. The paper reports on the effect of this approach on teacher pedagogy and on student learning of Ideas about Matter. The paper will present data from video of classroom activities, students’ work samples, student and teacher interviews and pre and post-unit testing, to explore what a representational focus might entail in teaching Ideas about Matter, and the role of representations in learning and reasoning and exploring scientific ideas.

Orbiting data: inciting centrifugal questions about embodied learning in science

In this paper we explore how reanimating a video data sequence with editing and creative software provided an opportunity for the data to speak and to demand new and surprising responses from us. Our data-ing brought new lines and spaces to the fore, through a process of refraction and re-animation which forced a focus on embodied inter-relationships and impeded precipitous analytical thought on the part of the researcher. We note how the aesthetic of the new images evoked awareness of our own part in the production of the object of our research. In particular, our own collegial interchange, punctuated by time and distance due to our respective locations on opposite sides of the globe, opened up a space for data-lingering in the intervening silences and pauses. Our choice of images engenders and reflects our sense of movement between the `I’ and the `we’ in their depiction of students’ learning about space.