Searcharxiv⌕ Search

arXiv subjects

Thomas Wong

Publications and source records attributed to Thomas Wong.

10 recordsLinked to original sources
arXiv2023

3D Printed Discrete Dielectric Lens With Improved Matching Layers

This paper presents a non-zoned discrete dielectric lens comprising two or three matching layers to reduce the 50-110 GHz frequency range reflections. Based on Chebyshev and binomial multi-section transformers, the designed models use matching layers at the top and bottom. In addition, the presented designs use pins instead of the conventional slots for the matching layers, thus easing the manufacturing process. The results show that the broadband realized gain obtained using the proposed design is higher for both the two- and three-layer design than the commonly used quarter-wave transformer. A Binomial lens with two matchings layers using 38 unit cells is fabricated and illuminated by an open-ended waveguide to validate the simulation results obtained using CST Microwave Studio. The fabrication process uses stereolithography additive manufacturing.

eess.SP↗
arXiv2023

Deep incremental learning models for financial temporal tabular datasets with distribution shifts

We present a robust deep incremental learning framework for regression tasks on financial temporal tabular datasets which is built upon the incremental use of commonly available tabular and time series prediction models to adapt to distributional shifts typical of financial datasets. The framework uses a simple basic building block (decision trees) to build self-similar models of any required complexity to deliver robust performance under adverse situations such as regime changes, fat-tailed distributions, and low signal-to-noise ratios. As a detailed study, we demonstrate our scheme using XGBoost models trained on the Numerai dataset and show that a two layer deep ensemble of XGBoost models over different model snapshots delivers high quality predictions under different market regimes. We also show that the performance of XGBoost models with different number of boosting rounds in three scenarios (small, standard and large) is monotonically increasing with respect to model size and converges towards the generalisation upper bound. We also evaluate the robustness of the model under variability of different hyperparameters, such as model complexity and data sampling settings. Our model has low hardware requirements as no specialised neural architectures are used and each base model can be independently trained in parallel.

cs.LG↗
arXiv2023

Online learning techniques for prediction of temporal tabular datasets with regime changes

The application of deep learning to non-stationary temporal datasets can lead to overfitted models that underperform under regime changes. In this work, we propose a modular machine learning pipeline for ranking predictions on temporal panel datasets which is robust under regime changes. The modularity of the pipeline allows the use of different models, including Gradient Boosting Decision Trees (GBDTs) and Neural Networks, with and without feature engineering. We evaluate our framework on financial data for stock portfolio prediction, and find that GBDT models with dropout display high performance, robustness and generalisability with reduced complexity and computational cost. We then demonstrate how online learning techniques, which require no retraining of models, can be used post-prediction to enhance the results. First, we show that dynamic feature projection improves robustness by reducing drawdown in regime changes. Second, we demonstrate that dynamical model ensembling based on selection of models with good recent performance leads to improved Sharpe and Calmar ratios of out-of-sample predictions. We also evaluate the robustness of our pipeline across different data splits and random seeds with good reproducibility.

q-fin.CP↗
arXiv2023

Feature Engineering Methods on Multivariate Time-Series Data for Financial Data Science Competitions

This paper is a work in progress. We are looking for collaborators to provide us financial datasets in Equity/Futures market to conduct more bench-marking studies. The authors have papers employing similar methods applied on the Numerai dataset, which is freely available but obfuscated. We apply different feature engineering methods for time-series to US market price data. The predictive power of models are tested against Numerai-Signals targets.

q-fin.ST↗
arXiv2021

Magnetic Nanohorns for measurement of the interfacial Dzyaloshinskii-Moriya interaction: A micromagnetic study

We evaluate the micromagnetic energy of a domain wall contained in a thin, horn-shaped nanostructure and propose a method to extract the strength of the interfacial Dzyaloshinskii-Moriya interaction based on the equilibrium position of the wall. An inplane magnetic field coupled with the interfacial DMI balances a competing driving force related to the "nanohorn" tapering and consequent variation in length. It is evident that the strength of the the DMI is directly correlated to the equilibrium position of the DW. Moreover, the range of measurable DMI values can be controlled by varying the degree of tapering in the nanohorn. Numerical and semi-analytical calculations were separately performed and show broad agreement. The calculations provided here offer an alternative framework for future measurement of DMI, which has been notoriously difficult to do experimentally.

cond-mat.mtrl-sci↗
arXiv2019

Semi-flexible directed polymers in a strip with attractive walls

We study a model of a semiflexible long chain polymer confined to a two-dimensional slit of width $w$, and interacting with the walls of the slit. The interactions with the walls are controlled by Boltzmann weights $a$ and $b$, and the flexibility of the polymer is controlled by another Boltzmann weight $c$. This is a simple model of the steric stabilisation of colloidal dispersions by polymers in solution. We solve the model exactly and compute various quantities in $(a,b,c)$-space, including the free energy and the force exerted by the polymer on the walls of the slit. In some cases these quantities can be computed exactly for all $w$, while for others only asymptotic expressions can be found. Of particular interest is the zero-force surface -- the manifold in $(a,b,c)$-space where the free energy is independent of $w$, and the loss of entropy due to confinement in the slit is exactly balanced by the energy gained from interactions with the walls.

cond-mat.stat-mech↗
arXiv2014

Confining multiple polymers between sticky walls: a directed walk model of two polymers

We study a model of two polymers confined to a slit with sticky walls. More precisely, we find and analyse the exact solution of two directed friendly walks in such a geometry on the square lattice. We compare the infinite slit limit, in which the length of the polymer (thermodynamic limit) is taken to infinity before the width of the slit is considered to become large, to the opposite situation where the order of the limits are swapped, known as the half-plane limit when one polymer is modelled. In contrast with the single polymer system we find that the half-plane and infinite slit limits coincide. We understand this result in part due to the tethering of polymers on both walls of the slit. We also analyse the entropic force exerted by the polymers on the walls of the slit. Again the results differ significantly from single polymer models. In a single polymer system both attractive and repulsive regimes were seen, whereas in our two walk model only repulsive forces are observed. We do, however, see that the range of the repulsive force is dependent on the parameter values. This variation can be explained by the adsorption of the walks on opposite walls of the slit.

cond-mat.stat-mech↗
arXiv2013

SOAP3-dp: Fast, Accurate and Sensitive GPU-based Short Read Aligner

To tackle the exponentially increasing throughput of Next-Generation Sequencing (NGS), most of the existing short-read aligners can be configured to favor speed in trade of accuracy and sensitivity. SOAP3-dp, through leveraging the computational power of both CPU and GPU with optimized algorithms, delivers high speed and sensitivity simultaneously. Compared with widely adopted aligners including BWA, Bowtie2, SeqAlto, GEM and GPU-based aligners including BarraCUDA and CUSHAW, SOAP3-dp is two to tens of times faster, while maintaining the highest sensitivity and lowest false discovery rate (FDR) on Illumina reads with different lengths. Transcending its predecessor SOAP3, which does not allow gapped alignment, SOAP3-dp by default tolerates alignment similarity as low as 60 percent. Real data evaluation using human genome demonstrates SOAP3-dp's power to enable more authentic variants and longer Indels to be discovered. Fosmid sequencing shows a 9.1 percent FDR on newly discovered deletions. SOAP3-dp natively supports BAM file format and provides a scoring scheme same as BWA, which enables it to be integrated into existing analysis pipelines. SOAP3-dp has been deployed on Amazon-EC2, NIH-Biowulf and Tianhe-1A.

q-bio.GN↗
arXiv2012

Exact solution of two friendly walks above a sticky wall with single and double interactions

We find, and analyse, the exact solution of two friendly directed walks, modelling polymers, which interact with a wall via contact interactions. We specifically consider two walks that begin and end together so as to imitate a polygon. We examine a general model in which a separate interaction parameter is assigned to configurations where both polymers touch the wall simultaneously, and investigate the effect this parameter has on the integrability of the problem. We find an exact solution of the generating function of the model, and provide a full analysis of the phase diagram that admits three phases with one first-order and two second-order transition lines between these phases. We argue that one physically realisable model would see two phase transitions as the temperature is lowered.

cond-mat.stat-mech↗