MRC Laboratory of Molecular Biology
CUED - Cambridge University Engineering DepartmentNot a member yet
45551 research outputs found
Sort by
Author Correction: Bacterial variability in the mammalian gut captured by a single-cell synthetic oscillator (Nature Communications, (2019), 10, 1, (4665), 10.1038/s41467-019-12638-z)
No full textIn the original version of the Supplementary Information provided with this Article, Supplementary Table 1 incorrectly provided the details for strain LPT239 as “E. coli MC4100 + pLPT234 + pLPT145” and for strain LPT320 as “E. coli MC4100 + pLPT234 + pLPT41”. The details for strain LPT239 should be “E. coli MC4100 + pLPT234 + pLPT41” and for strain LPT320 should be “E. coli MC4100 + pLPT234 + pLPT145”. This has been corrected in the HMTL version of the article and the correct table appended below: Original (Table presented.)
On Weakly Coupled Resonant MEMS Transducers Operating in the Modal Overlap Regime
No full textMiniaturized physical transducers based on weakly coupled resonators have previously demonstrated the twin benefits of high parametric sensitivity and the first-order common-mode rejection of environmental effects. Current approaches to sensing based on coupled resonator transducers employ strong coupling where the modal overlap of the responses is avoided. This strong coupling limits the sensitivity for such mode-localized sensors that utilize an amplitude ratio (AR) output metric as opposed to tracking resonant frequency shifts. In this article, this limitation is broken through by theoretically and experimentally demonstrating the operation of the weakly coupled resonators in the weak-coupling (modal overlap) regime. Especially, a prototype microelectromechanical systems (MEMS) sensor based on this principle is employed to detect shifts in stiffness, with a stiffness bias instability of 10.3\mu \text{N} /m (9.5 ppb) and a corresponding noise floor of 7.1\mu \text{N} /m/ \surd Hz (6.8 ppb/ \surd Hz). The linear dynamic range of such AR readout sensors is first explored and found to be defined by the dynamic range of the secondary resonator. The proposed method provides a promising approach for high-performance resonant force and inertial sensors
Superconducting Conductor on Round Core (CORC) Cables: 2D or 3D Modeling?
No full textThe HTSConductor on Round Core (CORC) Cables have been widely recognized as the proper candidate for superconducting applications such as the high-field magnets and power transmission cables. This article presents the modeling of HTS CORC cables using the FEM H-formulation model merged into the package COMSOL Multi-physics. The 2D and 3D H-formulation models of HTS CORC cables were compared and analyzed. There were two important parameters in our CORC design: (1) the gap angle between the HTS tapes in a single layer, and, (2) the degree of twist for HTS tapes (normally identified as the pitch). The 2D CORC model has much faster solving speed over the 3D model and sometimes is able to model the AC loss in a reasonable range, while, the 3D CORC model has the advantages over 2D CORC model with respect to more information in each sector with the different conditions of two parameters above, and more realistic loss calculation
A distributed optimisation framework combining natural gradient with Hessian-free for discriminative sequence training
No full textThis paper presents a novel natural gradient and Hessian-free (NGHF) optimisation framework for neural network training that can operate efficiently in a distributed manner. It relies on the linear conjugate gradient (CG) algorithm to combine the natural gradient (NG) method with local curvature information from Hessian-free (HF). A solution to a numerical issue in CG allows effective parameter updates to be generated with far fewer CG iterations than usually used (e.g. 5-8 instead of 200). This work also presents a novel preconditioning approach to improve the progress made by individual CG iterations for models with shared parameters. Although applicable to other training losses and model structures, NGHF is investigated in this paper for lattice-based discriminative sequence training for hybrid hidden Markov model acoustic models using a standard recurrent neural network, long short-term memory, and time delay neural network models for output probability calculation. Automatic speech recognition experiments are reported on the multi-genre broadcast data set for a range of different acoustic model types. These experiments show that NGHF achieves larger word error rate reductions than standard stochastic gradient descent or Adam, while requiring orders of magnitude fewer parameter updates
LCOS SLM based compact system of polarization modulation for data storage in glass
No full textThis paper introduces a compact dynamic polarization-modulation system using Liquid Crystal on Silicon Spatial Light Modulators (LCOS SLM) which can be used for data storage in glass. This new design uses two cascade-connected holograms on a single LCOS SLM in conjunction with a half-wave plate, to modulate multiple linear polarization states and simultaneously and generate desired holographic images. Furthermore, a zero-order suppression method is developed based on a computed holographic lens capable of on-axis and off-axis controls for the compact system to improve the quality of the data storage patterns generated by a laser light in glass
Factors influencing communication in collaborative design
No full textCommunication between people is commonly thought to play a major role in determining the success or failure of collaborative design projects. We conducted 8 weeks of observation and 62 interviews in three design companies. Our findings indicate that communication is blamed for causing many problems which are not of obvious technical nature. However, such problems are more often symptomatic of underlying factors related to information, representation, individual, team, and organisational issues. Examples of factors influencing communication at interfaces are people's understanding of the ‘sequence of tasks’ or ‘goals and objectives’ in the design process. With a system-theory informed conceptualisation of communication, we propose that awareness of such factors could assist design practitioners in targeting interventions and researchers in generating hypotheses about communication and design performance
Nested sampling for materials
No full textAbstract: We review the materials science applications of the nested sampling (NS) method, which was originally conceived for calculating the evidence in Bayesian inference. We describe how NS can be adapted to sample the potential energy surface (PES) of atomistic systems, providing a straightforward approximation for the partition function and allowing the evaluation of thermodynamic variables at arbitrary temperatures. After an overview of the basic method, we describe a number of extensions, including using variable cells for constant pressure sampling, the semi-grand-canonical approach for multicomponent systems, parallelizing the algorithm, and visualizing the results. We cover the range of materials applications of NS from the past decade, from exploring the PES of Lennard–Jones clusters to that of multicomponent condensed phase systems. We highlight examples how the information gained via NS promotes the understanding of materials properties through a novel way of visualizing the PES, identifying thermodynamically relevant basins, and calculating the entire pressure–temperature(–composition) phase diagram. Graphic abstract: [Figure not available: see fulltext.
A distributed robust state estimation algorithm for power systems considering maximum exponential absolute value
No full textIn traditional power system state estimation application, the distribution of measurement noise is frequently assumed to be Gaussian. However, the distribution is sometimes unknown in practice and the above assumption may be wrong. In this paper, a totally distributed robust state estimation algorithm considering maximum exponential absolute value (MEAV) is proposed for multi-area power systems. The state estimation optimization problem is founded on the basis of the maximum correntropy criterion with Laplace kernel function. Considering that the cost function of MEAV is nondifferentiable, the limited case of multiple segment (MS) estimator and the matrix-splitting technique are further utilized together to carry out the MEAV based algorithm within a fully distributed framework. Each local area only utilizes local information and limited information from neighboring areas but can obtain the estimation solution respectively. The communication load is alleviated and the robustness of state estimation is increased. The proposed distributed robust MEAV based algorithm has been tested using the IEEE 14-bus, 118-bus and 300-bus systems. The simulation results demonstrate the effectiveness and robustness of the proposed distributed robust MEAV based algorithm
Differential Thermal Conductivity CO2 Sensor
No full textIn this paper we present a novel MEMS technology for CO detection, based on differential thermal conductivity. Holes have been positioned across the thin-film dielectric membrane to induce an asymmetric temperature profile in conjunction with a centrally heated tungsten resistor, leading to distinct magnitudes of gas interaction across the sensor, all of which is substantiated by a numerical model. Differential readings are taken from four surrounding resistors and shown to correlate to CO concentration. The best resistor combination is investigated and is shown to provide a sensitivity of 0.05 mV/%CO. Using separate heating and sensing elements greatly lowers the noise and a detection resolution of < 100 ppm is calculated. This proof-of-concept article introduces a device that could lead to low-cost, low-power, high-sensitivity, scalable gas detection