• Canada
• Research softwareclose
• Other research productsclose
• Other ORP type close
• GB close

The phasic firing of dopamine neurons has been theorized to encode a reward-prediction error as formalized by the temporal-difference (TD) algorithm in reinforcement learning. Most TD models of dopamine have assumed a stimulus representation, known as the complete serial compound, in which each moment in a trial is distinctly represented. We introduce a more realistic temporal stimulus representation for the TD model. In our model, all external stimuli, including rewards, spawn a series of internal microstimuli, which grow weaker and more diffuse over time. These microstimuli are used by the TD learning algorithm to generate predictions of future reward. This new stimulus representation injects temporal generalization into the TD model and enhances correspondence between model and data in several experiments, including those when rewards are omitted or received early. This improved fit mostly derives from the absence of large negative errors in the new model, suggesting that dopamine alone can encode the full range of TD errors in these situations.

Digitalitzat per Artyplan

'Deconstructing voices' is a unique collaboration in sample culture, mixing live poetry with VJ, screen and digital technologies. The words are read live by the poets but have also been, and are during the performance, filmed and mixed (in a style influenced by the rave culture of the 80s and Russian Constructivism) with music and images. No two performances are ever the same but each is a unique way of performing poetry which honours the language and intention of serious poets but also engages the audience in a vital and imaginative way. The performance brings together a unique blend of expertise from established artists and performers working alongside emerging creative talent.

From social sciences to biology and physics, gamified systems and games are increasingly being used as contexts and tools for research: as "petri dishes" for observing macro-social and economic dynamics; as sources of "big" and/or ecologically valid user behavior and health data; as crowdsourcing tools for research tasks; or as a means to motivate e.g. survey completion. However, this gamification of research comes with significant ethical ramifications. This workshop therefore explores opportunities, challenges, best practices, and ethical issues arising from different strategies of gamifying research.

Part of the Roy Newsome Brass Band Archive

Use of document genre in information retrieval systems has the potential to improve the task-appropriateness of results. However, genre classification remains a challenging problem. We describe a case study of genre classification in a software engineering workplace domain, which includes the development of a genre taxonomy and experiments in automatic genre classification using supervised machine learning. We present results based on evaluation using real-life enterprise data from this work domain.

The complex nature of human cognition has resulted in cognitive genomics lagging behind many other fields in terms of gene discovery using genome-wide association study (GWAS) methods. In an attempt to overcome these barriers, the current study utilized GWAS meta-analysis to examine the association of common genetic variation (~8M single-nucleotide polymorphisms (SNP) with minor allele frequency ⩾1%) to general cognitive function in a sample of 35 298 healthy individuals of European ancestry across 24 cohorts in the Cognitive Genomics Consortium (COGENT). In addition, we utilized individual SNP lookups and polygenic score analyses to identify genetic overlap with other relevant neurobehavioral phenotypes. Our primary GWAS meta-analysis identified two novel SNP loci (top SNPs: rs76114856 in the CENPO gene on chromosome 2 and rs6669072 near LOC105378853 on chromosome 1) associated with cognitive performance at the genome-wide significance level (P<5 × 10^−8). Gene-based analysis identified an additional three Bonferroni-corrected significant loci at chromosomes 17q21.31, 17p13.1 and 1p13.3. Altogether, common variation across the genome resulted in a conservatively estimated SNP heritability of 21.5% (s.e.=0.01%) for general cognitive function. Integration with prior GWAS of cognitive performance and educational attainment yielded several additional significant loci. Finally, we found robust polygenic correlations between cognitive performance and educational attainment, several psychiatric disorders, birth length/weight and smoking behavior, as well as a novel genetic association to the personality trait of openness. These data provide new insight into the genetics of neurocognitive function with relevance to understanding the pathophysiology of neuropsychiatric illness.

Enmeshed 3, for cello and live electronics, is the third in a series of works in which a solo instrument becomes ‘enmeshed’ in multiple layers of transformations derived from the live performance. The works are shaped and structured in terms of the varying relationships between these layers and the ‘distances’ between the original acoustic performance and the various transformations, in terms of pitch, time delay, timbre, texture and space. At certain points in the work these almost converge whilst at other times large distances open up, with the different layers in a wild counterpoint. All the sounds in the work derive from live transformation of the soloist's performance. The composer’s own granular synthesis algorithms play a significant role in these transformations. \ud \ud Multichannel spatialisation also plays an important part in terms of spatial positioning and movement, the creation of different virtual spatial environments and in the definition of different layers. It can be performed variously with between 8 and 24 channels. Enmeshed 3 is in five contrasting but inter-related sections centering around a long slow meditative central passage. It was written for Madeleine Shapiro who premiered it at the New York City Electroacoustic Music Festival in April 2013.

Aquistore is a deep saline CO2 storage research and demonstration project located near Estevan, Saskatchewan where CO2 is transported via pipeline and injected into a sandstone reservoir ∼3200 m below the surface. A pre-injection time-lapse analysis performed on two sparse 3D seismic datasets was used to characterise the background time-lapse signal-to-noise level at the storage site. The time-lapse analysis revealed that the lowest global nRMS was 0.07 which was taken to represent the level above which CO2 would be detectable in the reservoir. We investigate the conditions under which the injected CO2 can be detected above the defined minimum noise level through Gassmann fluid substitution and 3D seismic forward modelling. Additionally, Wave Unix was used to simulate the seismic response of the reservoir due to the injected CO2 by generating the synthetic surface reflection seismic data from an explosive surface P-wave source. We generated noise-free synthetic seismograms for the baseline model as well as for the 2-phase fluid replacement of brine with CO2 for CO2 concentrations up to 100% within the target zone – the monitors. The baseline and monitor traces from the 3D seismic survey at Aquistore are used as the noise traces in this study, and were added to their respective baseline and monitor synthetic traces. The nRMS within the reservoir was then computed for the noisy baseline and various noisy monitor surveys and was used in the assessment of the limitation to the detection of the injected CO2 in the reservoir under the background noise level at the site. We are able to conclude that the time-lapse repeatability will not limit the ability to monitor the CO2 induced changes in the reservoir at the Aquistore storage site.