| 1. |
- Thoma, B, et al.
(författare)
-
An international, interprofessional investigation of the self-reported podcast listening habits of emergency clinicians: A METRIQ Study
- 2020
-
Ingår i: CJEM. - : Springer Science and Business Media LLC. - 1481-8043 .- 1481-8035. ; 22:1, s. 112-117
-
Tidskriftsartikel (refereegranskat)abstract
- ObjectivesPodcasts are increasingly being used for medical education. A deeper understanding of usage patterns would inform both producers and researchers of medical podcasts. We aimed to determine how and why podcasts are used by emergency medicine and critical care clinicians.MethodsAn international interprofessional sample (medical students, residents, physicians, nurses, physician assistants, and paramedics) was recruited through direct contact and a multimodal social media (Twitter and Facebook) campaign. Each participant completed a survey outlining how and why they utilize medical podcasts. Recruitment materials included an infographic and study website.Results390 participants from 33 countries and 4 professions (medicine, nursing, paramedicine, physician assistant) completed the survey. Participants most frequently listened to medical podcasts to review new literature (75.8%), learn core material (75.1%), and refresh memory (71.8%). The majority (62.6%) were aware of the ability to listen at increased speeds, but most (76.9%) listened at 1.0 x (normal) speed. All but 25 (6.4%) participants concurrently performed other tasks while listening. Driving (72.3%), exercising (39.7%), and completing chores (39.2%) were the most common. A minority of participants used active learning techniques such as pausing, rewinding, and replaying segments of the podcast. Very few listened to podcasts multiple times.ConclusionsAn international cohort of emergency clinicians use medical podcasts predominantly for learning. Their listening habits (rarely employing active learning strategies and frequently performing concurrent tasks) may not support this goal. Further exploration of the impact of these activities on learning from podcasts is warranted.
|
|
| 2. |
- Valiente-Dobon, J. J., et al.
(författare)
-
Conceptual design of the AGATA 2 pi array at LNL
- 2023
-
Ingår i: Nuclear Instruments and Methods in Physics Research Section A. - : Elsevier BV. - 0168-9002 .- 1872-9576. ; 1049
-
Tidskriftsartikel (refereegranskat)abstract
- The Advanced GAmma Tracking Array (AGATA) has been installed at Laboratori Nazionali di Legnaro (LNL), Italy. In this installation, AGATA will consist, at the beginning, of 13 AGATA triple clusters (ATCs) with an angular coverage of 1n,and progressively the number of ATCs will increase up to a 2 pi angular coverage. This setup will exploit both stable and radioactive ion beams delivered by the Tandem-PIAVE-ALPI accelerator complex and the SPES facility. The new implementation of AGATA at LNL will be used in two different configurations, firstly one coupled to the PRISMA large-acceptance magnetic spectrometer and lately a second one at Zero Degrees, along the beam line. These two configurations will allow us to cover a broad physics program, using different reaction mechanisms, such as Coulomb excitation, fusion-evaporation, transfer and fission at energies close to the Coulomb barrier. These setups have been designed to be coupled with a large variety of complementary detectors such as charged particle detectors, neutron detectors, heavy-ion detectors, high-energy gamma-ray arrays, cryogenic and gasjet targets and the plunger device for lifetime measurements. We present in this paper the conceptual design, characteristics and performance figures of this implementation of AGATA at LNL.
|
|
| 3. |
|
|
| 4. |
|
|
| 5. |
- Cerutti, G., et al.
(författare)
-
Sound event detection with binary neural networks on tightly power-constrained IoT devices
- 2020
-
Ingår i: ACM International Conference Proceeding Series. - New York, NY, USA : Association for Computing Machinery. - 9781450370530
-
Konferensbidrag (refereegranskat)abstract
- Sound event detection (SED) is a hot topic in consumer and smart city applications. Existing approaches based on deep neural networks (DNNs) are very effective, but highly demanding in terms of memory, power, and throughput when targeting ultra-low power always-on devices. Latency, availability, cost, and privacy requirements are pushing recent IoT systems to process the data on the node, close to the sensor, with a very limited energy supply, and tight constraints on the memory size and processing capabilities precluding to run state-of-The-Art DNNs. In this paper, we explore the combination of extreme quantization to a small-footprint binary neural network (BNN) with the highly energy-efficient, RISC-V-based (8+1)-core GAP8 microcontroller. Starting from an existing CNN for SED whose footprint (815 kB) exceeds the 512 kB of memory available on our platform, we retrain the network using binary filters and activations to match these memory constraints. (Fully) binary neural networks come with a natural drop in accuracy of 12-18% on the challenging ImageNet object recognition challenge compared to their equivalent full-precision baselines. This BNN reaches a 77.9% accuracy, just 7% lower than the full-precision version, with 58 kB (7.2× less) for the weights and 262 kB (2.4× less) memory in total. With our BNN implementation, we reach a peak throughput of 4.6 GMAC/s and 1.5 GMAC/s over the full network, including preprocessing with Mel bins, which corresponds to an efficiency of 67.1 GMAC/s/W and 31.3 GMAC/s/W, respectively. Compared to the performance of an ARM Cortex-M4 implementation, our system has a 10.3× faster execution time and a 51.1× higher energy-efficiency. © 2020 ACM.
|
|
