Searcharxiv⌕ Search

arXiv subjects

David Hill

Publications and source records attributed to David Hill.

At least 19 recordsLinked to original sources
arXiv2025

A K-Means, Ward and DBSCAN repeatability study

Reproducibility is essential in machine learning because it ensures that a model or experiment yields the same scientific conclusion. For specific algorithms repeatability with bitwise identical results is also a key for scientific integrity because it allows debugging. We decomposed several very popular clustering algorithms: K-Means, DBSCAN and Ward into their fundamental steps, and we identify the conditions required to achieve repeatability at each stage. We use an implementation example with the Python library scikit-learn to examine the repeatable aspects of each method. Our results reveal inconsistent results with K-Means when the number of OpenMP threads exceeds two. This work aims to raise awareness of this issue among both users and developers, encouraging further investigation and potential fixes.

cs.LG↗
arXiv2025

Scaling Fabric-Based Piezoresistive Sensor Arrays for Whole-Body Tactile Sensing

Scaling tactile sensing for robust whole-body manipulation is a significant challenge, often limited by wiring complexity, data throughput, and system reliability. This paper presents a complete architecture designed to overcome these barriers. Our approach pairs open-source, fabric-based sensors with custom readout electronics that reduce signal crosstalk to less than 3.3% through hardware-based mitigation. Critically, we introduce a novel, daisy-chained SPI bus topology that avoids the practical limitations of common wireless protocols and the prohibitive wiring complexity of USB hub-based systems. This architecture streams synchronized data from over 8,000 taxels across 1 square meter of sensing area at update rates exceeding 50 FPS, confirming its suitability for real-time control. We validate the system's efficacy in a whole-body grasping task where, without feedback, the robot's open-loop trajectory results in an uncontrolled application of force that slowly crushes a deformable cardboard box. With real-time tactile feedback, the robot transforms this motion into a gentle, stable grasp, successfully manipulating the object without causing structural damage. This work provides a robust and well-characterized platform to enable future research in advanced whole-body control and physical human-robot interaction.

cs.RO↗
arXiv2025

Molecular adhesion assay for biopolymer systems

Molecular adhesion plays a central role in many biological systems, yet existing methods to quantify adhesive strength often struggle to bridge the gap between single-molecule resolution and biologically relevant environments. Here, we present a scalable micromagnetic bead-based adhesion assay capable of quantifying detachment forces under physiologically meaningful conditions. Designed to probe mucoadhesion in the context of mucociliary clearance, our system applies controlled magnetic forces to ligand-coated beads adhered to functionalized substrates and tracks detachment events using high-speed microscopy and calibrated z-displacement mapping. The platform combines substrate- and bead-side surface chemistry control with high-throughput imaging and in situ force calibration via Stokes drag. We demonstrate the ability to distinguish sub-nanonewton to nanonewton force regimes across a range of bead-substrate pairings, including COOH-COOH, PEG-PEG, and cell culture-derived human bronchial epithelial (HBE) mucus interactions. Surface functionalization was validated via fluorescence imaging and zeta potential measurements, while detachment forces were used to estimate binding energy and infer dissociation constants. This assay enables detailed characterization of multivalent, force-sensitive adhesive interactions and offers a powerful new approach for studying bioadhesive systems, including mucus-pathogen interactions and drug delivery materials.

physics.bio-ph↗
arXiv2024

Improving SSVEP BCI Spellers With Data Augmentation and Language Models

Steady-State Visual Evoked Potential (SSVEP) spellers are a promising communication tool for individuals with disabilities. This Brain-Computer Interface utilizes scalp potential data from (electroencephalography) EEG electrodes on a subject's head to decode specific letters or arbitrary targets the subject is looking at on a screen. However, deep neural networks for SSVEP spellers often suffer from low accuracy and poor generalizability to unseen subjects, largely due to the high variability in EEG data. In this study, we propose a hybrid approach combining data augmentation and language modeling to enhance the performance of SSVEP spellers. Using the Benchmark dataset from Tsinghua University, we explore various data augmentation techniques, including frequency masking, time masking, and noise injection, to improve the robustness of deep learning models. Additionally, we integrate a language model (CharRNN) with EEGNet to incorporate linguistic context, significantly enhancing word-level decoding accuracy. Our results demonstrate accuracy improvements of up to 2.9 percent over the baseline, with time masking and language modeling showing the most promise. This work paves the way for more accurate and generalizable SSVEP speller systems, offering improved communication solutions for individuals with disabilities.

cs.HC↗
arXiv2016

Quantum supergroups V. Braid group action

We construct a braid group action on quantum covering groups. We further use this action to construct a PBW basis for the positive half in finite type which is pairwise-orthogonal under the inner product. This braid group action is induced by operators on the integrable modules; however, these operators satisfy spin braid relations.

math.QA↗
arXiv2015

Reliable Initialization of GPU-enabled Parallel Stochastic Simulations Using Mersenne Twister for Graphics Processors

Parallel stochastic simulations tend to exploit more and more computing power and they are now also developed for General Purpose Graphics Process Units (GP-GPUs). Conse-quently, they need reliable random sources to feed their applications. We propose a survey of the current Pseudo Random Numbers Generators (PRNG) available on GPU. We give a particular focus to the recent Mersenne Twister for Graphics Processors (MTGP) that has just been released. Our work provides empirically checked statuses designed to initialize a particular configuration of this generator, in order to prevent any potential bias introduced by the parallelization of the PRNG.

cs.DC↗
arXiv2015

Warp-Level Parallelism: Enabling Multiple Replications In Parallel on GPU

Stochastic simulations need multiple replications in order to build confidence intervals for their results. Even if we do not need a large amount of replications, it is a good practice to speed-up the whole simulation time using the Multiple Replications In Parallel (MRIP) approach. This approach usually supposes to have access to a parallel computer such as a symmetric mul-tiprocessing machine (with many cores), a computing cluster or a computing grid. In this paper, we propose Warp-Level Parallelism (WLP), a GP-GPU-enabled solution to compute MRIP on GP-GPUs (General-Purpose Graphics Processing Units). These devices display a great amount of parallel computational power at low cost, but are tuned to process efficiently the same operation on several data, through different threads. Indeed, this paradigm is called Single Instruction, Multiple Threads (SIMT). Our approach proposes to rely on small threads groups, called warps, to perform independent computations such as replications. We have benchmarked WLP with three different models: it allows MRIP to be computed up to six times faster than with the SIMT computing paradigm.

cs.DC↗
arXiv2014

How to Correctly Deal With Pseudorandom Numbers in Manycore Environments - Application to GPU programming with Shoverand

Stochastic simulations are often sensitive to the source of randomness that character-izes the statistical quality of their results. Consequently, we need highly reliable Random Number Generators (RNGs) to feed such applications. Recent developments try to shrink the computa-tion time by relying more and more General Purpose Graphics Processing Units (GP-GPUs) to speed-up stochastic simulations. Such devices bring new parallelization possibilities, but they also introduce new programming difficulties. Since RNGs are at the base of any stochastic simulation, they also need to be ported to GP-GPU. There is still a lack of well-designed implementations of quality-proven RNGs on GP-GPU platforms. In this paper, we introduce ShoveRand, a frame-work defining common rules to generate random numbers uniformly on GP-GPU. Our framework is designed to cope with any GPU-enabled development platform and to expose a straightfor-ward interface to users. We also provide an existing RNG implementation with this framework to demonstrate its efficiency in both development and ease of use.

cs.DC↗
arXiv2014

Quantum Shuffles and Quantum Supergroups of Basic Type

We initiate the study of several distinguished bases for the positive half of a quantum supergroup $U_q$ associated to a general super Cartan datum $(\mathrm{I}, (\cdot,\cdot))$ of basic type inside a quantum shuffle superalgebra. The combinatorics of words for an arbitrary total ordering on $\mathrm{I}$ is developed in connection with the root system associated to $\mathrm{I}$. The monomial, Lyndon, and PBW bases of $U_q$ are constructed, and moreover, a direct proof of the orthogonality of the PBW basis is provided within the framework of quantum shuffles. Consequently, the canonical basis is constructed for $U_q$ associated to the standard super Cartan datum of type $\mathfrak{gl}(n|1)$, $\mathfrak{osp}(1|2n)$, or $\mathfrak{osp}(2|2n)$ or an arbitrary non-super Cartan datum. In the non-super case, this refines Leclerc's work and provides a new self-contained construction of canonical bases. The canonical bases of $U_q$, of its polynomial modules, as well as of Kac modules in the case of quantum $\mathfrak{gl}(2|1)$ are explicitly worked out.

math.QA↗
arXiv2013

Quantum supergroups II. Canonical basis

Following Kashiwara's algebraic approach, we construct crystal bases and canonical bases for quantum supergroups with no isotropic odd roots and for their integrable modules.

math.QA↗
arXiv2013

Quantum Supergroups I. Foundations

In this part one of a series of papers, we introduce a new version of quantum covering and super groups with no isotropic odd simple root, which is suitable for the studies of integrable modules, integral forms and bar-involution. A quantum covering group involves a quantum parameter q and a sign parameter pi squaring to 1, and it specializes to a quantum supergroup when pi=-1. Following Lusztig, we formulate and establish various structural results of the quantum covering groups, including bilinear form, quasi-R-matrix, Casimir, character formulas for integrable modules, and higher Serre relations.

math.QA↗
arXiv2013

Categorification of quantum Kac-Moody superalgebras

We introduce a non-degenerate bilinear form and use it to provide a new characterization of quantum Kac-Moody superalgebras with no isotropic odd simple roots. We show that the spin quiver Hecke algebras introduced by Kang-Kashiwara-Tsuchioka provide a categorification of half the quantum Kac-Moody superalgebras, using the recent work of Ellis-Khovanov-Lauda. A new idea here is that a super sign is categorified as spin (i.e., the parity-shift functor).

math.QA↗
arXiv2009

Representations of Quiver Hecke Algebras via Lyndon Bases

A new class of algebras have been introduced by Khovanov and Lauda and independently by Rouquier. These algebras categorify one-half of the Quantum group associated to arbitrary Cartan data. In this paper, we use the combinatorics of Lyndon words to construct the irreducible representations of those algebras associated to Cartan data of finite type. This completes the classification of simple modules for the quiver Hecke algebra initiated by Kleshchev and Ram.

math.RT↗
arXiv2009

Degenerate Affine Hecke-Clifford Algebras and Type $Q$ Lie Superalgebras

We construct the finite dimensional simple integral modules for the (degenerate) affine Hecke-Clifford algebra (AHCA). Our construction includes an analogue of Zelevinsky's segment representations, a complete combinatorial description of the simple calibrated modules, and a classification of the simple integral modules. Additionally, we construct an analogue of the Arakawa-Suzuki functor for the Lie superalgebra of type Q.

math.RT↗
arXiv2008

Cartan Invariants of Symmetric Groups and Iwahori-Hecke Algebras

Külshammer, Olsson and Robinson conjectured that a certain set of numbers determined the invariant factors of the $\ell$-Cartan matrix for $S_n$ (equivalently, the invariant factors of the Cartan matrix for the Iwahori-Hecke algebra $\mathcal{H}_n(q)$, where $q$ is a primitive $\ell$th root of unity). We call these invariant factors Cartan invariants. In a previous paper, the second author calculated these Cartan invariants when $\ell=p^r$, $p$ prime, and $r\leq p$ and went on to conjecture that the formulae should hold for all $r$. Another result was obtained, which is surprising and counterintuitive from a block theoretic point of view. Namely, given the prime decomposition $\ell=p_1^{r_1}... p_k^{r_k}$, the Cartan matrix of an $\ell$-block of $S_n$ is a product of Cartan matrices associated to $p_i^{r_i}$-blocks of $S_n$. In particular, the invariant factors of the Cartan matrix associated to an $\ell$-block of $S_n$ can be recovered from the Cartan matrices associated to the $p_i^{r_i}$-blocks. In this paper, we formulate an explicit combinatorial determination of the Cartan invariants of $S_n$--not only for the full Cartan matrix, \emph{but for an individual block}. We collect evidence for this conjecture, by showing that the formulae predict the correct determinant of the $\ell$-Cartan matrix. We then go on to show that Hill's conjecture implies the conjecture of KOR.

math.GR↗
arXiv2008

GoArrays: highly dynamic and efficient microarray probe design

MOTIVATION: The use of oligonucleotide microarray technology requires a very detailed attention to the design of specific probes spotted on the solid phase. These problems are far from being commonplace since they refer to complex physicochemical constraints. Whereas there are more and more publicly available programs for microarray oligonucleotide design, most of them use the same algorithm or criteria to design oligos, with only little variation. RESULTS: We show that classical approaches used in oligo design software may be inefficient under certain experimental conditions, especially when dealing with complex target mixtures. Indeed, our biological model is a human obligate parasite, the microsporidia Encephalitozoon cuniculi. Targets that are extracted from biological samples are composed of a mixture of pathogen transcripts and host cell transcripts. We propose a new approach to design oligonucleotides which combines good specificity with a potentially high sensitivity. This approach is original in the biological point of view as well as in the algorithmic point of view. We also present an experimental validation of this new strategy by comparing results obtained with standard oligos and with our composite oligos. A specific E.cuniculi microarray will overcome the difficulty to discriminate the parasite mRNAs from the host cell mRNAs demonstrating the power of the microarray approach to elucidate the lifestyle of an intracellular pathogen using mix mRNAs.

q-bio.QM↗