arXiv subjects

Nathan Cooper

Publications and source records attributed to Nathan Cooper.

At least 19 recordsLinked to original sources

arXiv2026

Enhancing Code Understanding for Impact Analysis by Combining Transformers and Program Dependence Graphs

Impact analysis (IA) is a critical software maintenance task that identifies the effects of a given set of code changes on a larger software project with the intention of avoiding potential adverse effects. IA is a cognitively challenging task that involves reasoning about the abstract relationships between various code constructs. Given its difficulty, researchers have worked to automate IA with approaches that primarily use coupling metrics as a measure of the "connectedness" of different parts of a software project. Many of these coupling metrics rely on static, dynamic, or evolutionary information and are based on heuristics that tend to be brittle, require expensive execution analysis, or large histories of co-changes to accurately estimate impact sets. In this paper, we introduce a novel IA approach, called Athena, that combines a software system's dependence graph information with a conceptual coupling approach that uses advances in deep representation learning for code without the need for change histories and execution information. Previous IA benchmarks are small, containing fewer than ten software projects, and suffer from tangled commits, making it difficult to measure accurate results. Therefore, we constructed a large-scale IA benchmark, called Alexandria, from 25 open-source software projects, that utilizes fine-grained commit information from bug fixes. On this new benchmark, our best-performing approach configuration achieves mRR, mAP, and HIT@10 scores of 60.32%, 35.19%, and 81.48%, respectively. Through various ablations and qualitative analyses, we show that Athena's novel combination of program dependence graphs and conceptual coupling information leads it to outperform a simpler baseline by 10.34%, 9.55%, and 11.68% with statistical significance.

cs.SE↗

arXiv2026

Enhanced Atom Capture via Multi-Frequency Magneto-Optical Trapping

Magneto-optical traps are central to atomic and molecular quantum technologies and precision tests of fundamental physics, where both sensitivity and bandwidth scale strongly with atom number and loading rate. We demonstrate that employing multiple, closely spaced optical frequency components in the cooling light of a $^{87}$Rb magneto-optical trap -- without utilizing any additional slowing techniques -- can double the steady state atom number and increase the loading rate by up to a factor of 4, compared to a conventional single-frequency implementation. Subsequently, we capture up to $1.0(1)\times10^{10}$ atoms with a loading rate of up to $1.3(2)\times 10^{11}\,\mathrm{atoms\,s^{-1}}$ from a thermal background. Numerical simulations reproduce the observed trends and predict substantially larger gains for increased trap sizes beyond our experimental bounds. By re-examining earlier studies of multi-frequency atom capture in the context of modern experimental hardware and emerging applications, we show that previously identified limitations can be avoided and establish multi-frequency cooling as a practical and scalable route to high-flux cold-atom sources. These results have immediate applications in portable atom-based quantum sensing, where higher bandwidth and precision can be achieved without forgoing compactness, and in atom-interferometric tests of fundamental physics, which benefit from access to larger-mass quantum systems.

physics.atom-ph↗

arXiv2025

Exploiting complex 3D-printed surface structures for portable quantum technologies

Portable quantum technologies require robust, lightweight apparatus with superior performance. For techniques dependent upon high-vacuum environments, such as atom interferometers and atomic clocks, 3D-printing enables new avenues to tailor in-vacuum gas propagation dynamics. We demonstrate intricate, fine-scale surface patterning of 3D-printed vacuum components to increase the rate at which gas particles collide with the surface. By applying a non-evaporable getter coating for use as a surface pump, we show that the patterned surface pumps gas particles 3.8 times faster than an equivalent flat areas. These patterns can be directly integrated into additively manufactured components, enabling application in close proximity to key experimental regions and contributing to overall mass-reduction. We develop numerical simulations that show good agreement with this result and predict up to a ten-fold increase in pumping rate, for realistic surface structures. Our work has direct applications in enabling passively-pumped portable quantum technologies, but also establishes 3D-printing as a powerful technique for the creation of optimized surface patterning to provide enhanced control over high-vacuum gas dynamics for a broad range of applications.

quant-ph↗

arXiv2025

It's All in The [MASK]: Simple Instruction-Tuning Enables BERT-like Masked Language Models As Generative Classifiers

While encoder-only models such as BERT and ModernBERT are ubiquitous in real-world NLP applications, their conventional reliance on task-specific classification heads can limit their applicability compared to decoder-based large language models (LLMs). In this work, we introduce ModernBERT-Large-Instruct, a 0.4B-parameter encoder model that leverages its masked language modelling (MLM) head for generative classification. Our approach employs an intentionally simple training loop and inference mechanism that requires no heavy pre-processing, heavily engineered prompting, or architectural modifications. ModernBERT-Large-Instruct exhibits strong zero-shot performance on both classification and knowledge-based tasks, outperforming similarly sized LLMs on MMLU and achieving 93% of Llama3-1B's MMLU performance with 60% less parameters. We also demonstrate that, when fine-tuned, the generative approach using the MLM head matches or even surpasses traditional classification-head methods across diverse NLU tasks.This capability emerges specifically in models trained on contemporary, diverse data mixes, with models trained on lower volume, less-diverse data yielding considerably weaker performance. Although preliminary, these results demonstrate the potential of using the original generative masked language modelling head over traditional task-specific heads for downstream tasks. Our work suggests that further exploration into this area is warranted, highlighting many avenues for future improvements.

cs.CL↗

arXiv2025

On the Generalizability of Transformer Models to Code Completions of Different Lengths

The programming landscape is nowadays being reshaped by the advent of Large Language Models (LLMs) able to automate code-related tasks related to code implementation (e.g., code completion) and comprehension (e.g., code summarization). Such a paradigm shift comes with a number of implications related to how software will be written, maintained, and evolved. Also, these LLMs are extremely expensive to train, posing questions on their sustainability over time. Given their training cost, their ability to generalize, namely their ability to work on task instances different from those on which they have been trained, is an aspect worth being investigated. Previous work already showed that transformer models can successfully support code completion in a cross-project setting. However, it is unclear whether LLM are able to generalize to inputs having lengths not seen during training. For example, it is known that training a model on short instances allows to substantially reduce the training cost. However, the extent to which such a model would provide good performance on sequences having lengths not seen during training is not known. Many recent works in Natural Language Processing (NLP) tackled this problem in the context of decoder-only LLMs, i.e., xPOS and ALiBi. To assess if these solutions extend to encoder-decoder LLMs usually adopted in the code-related tasks, we present a large empirical study evaluating this generalization property of these and other encoding schemes proposed in the literature, namely Sinusoidal, xPOS, ALiBi, and T5. We found that none of these solutions successfully generalize to unseen lengths and that the only safe solution is to ensure the representativeness in the training set of all lengths likely to be encountered at inference time.

cs.SE↗

arXiv2024

Smarter, Better, Faster, Longer: A Modern Bidirectional Encoder for Fast, Memory Efficient, and Long Context Finetuning and Inference

Encoder-only transformer models such as BERT offer a great performance-size tradeoff for retrieval and classification tasks with respect to larger decoder-only models. Despite being the workhorse of numerous production pipelines, there have been limited Pareto improvements to BERT since its release. In this paper, we introduce ModernBERT, bringing modern model optimizations to encoder-only models and representing a major Pareto improvement over older encoders. Trained on 2 trillion tokens with a native 8192 sequence length, ModernBERT models exhibit state-of-the-art results on a large pool of evaluations encompassing diverse classification tasks and both single and multi-vector retrieval on different domains (including code). In addition to strong downstream performance, ModernBERT is also the most speed and memory efficient encoder and is designed for inference on common GPUs.

cs.CL↗

arXiv2024

Rephrasing natural text data with different languages and quality levels for Large Language Model pre-training

Recently published work on rephrasing natural text data for pre-training LLMs has shown promising results when combining the original dataset with the synthetically rephrased data. We build upon previous work by replicating existing results on C4 and extending them with our optimized rephrasing pipeline to the English, German, Italian, and Spanish Oscar subsets of CulturaX. Our pipeline leads to increased performance on standard evaluation benchmarks in both the mono- and multilingual setup. In addition, we provide a detailed study of our pipeline, investigating the choice of the base dataset and LLM for the rephrasing, as well as the relationship between the model size and the performance after pre-training. By exploring data with different perceived quality levels, we show that gains decrease with higher quality. Furthermore, we find the difference in performance between model families to be bigger than between different model sizes. This highlights the necessity for detailed tests before choosing an LLM to rephrase large amounts of data. Moreover, we investigate the effect of pre-training with synthetic data on supervised fine-tuning. Here, we find increasing but inconclusive results that highly depend on the used benchmark. These results (again) highlight the need for better benchmarking setups. In summary, we show that rephrasing multilingual and low-quality data is a very promising direction to extend LLM pre-training data.

cs.CL↗

arXiv2024

Adaptive, symmetry-informed Bayesian metrology for precise quantum technology measurements

High precision measurements are essential to solve major scientific and technological challenges, from gravitational wave detection to healthcare diagnostics. Quantum sensing delivers greater precision, but an in-depth optimisation of measurement procedures has been overlooked. Here we present a systematic strategy for parameter estimation in the low-data limit that integrates experimental control parameters and natural symmetries. The method is guided by a Bayesian quantifier of precision gain, enabling adaptive optimisation tailored to the experiment. We provide general expressions for optimal estimators for any parameter. The strategy's power is demonstrated in a quantum technology experiment, in which ultracold caesium atoms are confined in a micromachined hole in an optical fibre. We find a five-fold reduction in the fractional variance of the estimated parameter, compared to the standard measurement procedure. Equivalently, our strategy achieves a target precision with a third of the data points previously required. Such enhanced device performance and accelerated data collection will be essential for applications in quantum computing, communication, metrology, and the wider quantum technology sector.

quant-ph↗

arXiv2024

Semantic GUI Scene Learning and Video Alignment for Detecting Duplicate Video-based Bug Reports

Video-based bug reports are increasingly being used to document bugs for programs centered around a graphical user interface (GUI). However, developing automated techniques to manage video-based reports is challenging as it requires identifying and understanding often nuanced visual patterns that capture key information about a reported bug. In this paper, we aim to overcome these challenges by advancing the bug report management task of duplicate detection for video-based reports. To this end, we introduce a new approach, called JANUS, that adapts the scene-learning capabilities of vision transformers to capture subtle visual and textual patterns that manifest on app UI screens - which is key to differentiating between similar screens for accurate duplicate report detection. JANUS also makes use of a video alignment technique capable of adaptive weighting of video frames to account for typical bug manifestation patterns. In a comprehensive evaluation on a benchmark containing 7,290 duplicate detection tasks derived from 270 video-based bug reports from 90 Android app bugs, the best configuration of our approach achieves an overall mRR/mAP of 89.8%/84.7%, and for the large majority of duplicate detection tasks, outperforms prior work by around 9% to a statistically significant degree. Finally, we qualitatively illustrate how the scene-learning capabilities provided by Janus benefits its performance.

cs.SE↗

arXiv2024

Additive Manufacturing of functionalised atomic vapour cells for next-generation quantum technologies

Atomic vapour cells are an indispensable tool for quantum technologies (QT), but potential improvements are limited by the capacities of conventional manufacturing methods. Using an additive manufacturing (AM) technique - vat polymerisation by digital light processing - we demonstrate, for the first time, a 3D-printed glass vapour cell. The exploitation of AM capacities allows intricate internal architectures, overprinting of 2D optoelectronical materials to create integrated sensors and surface functionalisation, while also showing the ability to tailor the optical properties of the AM glass by in-situ growth of gold nanoparticles. The produced cells achieve ultra-high vacuum of $2 \times 10^{-9}$ mbar and enable Doppler-free spectroscopy; we demonstrate laser frequency stabilisation as a QT application. These results highlight the transformative role that AM can play for QT in enabling compact, optimised and integrated multi-material components and devices.

physics.atom-ph↗

arXiv2024

Perplexed: Understanding When Large Language Models are Confused

Large Language Models (LLMs) have become dominant in the Natural Language Processing (NLP) field causing a huge surge in progress in a short amount of time. However, their limitations are still a mystery and have primarily been explored through tailored datasets to analyze a specific human-level skill such as negation, name resolution, etc. In this paper, we introduce perplexed, a library for exploring where a particular language model is perplexed. To show the flexibility and types of insights that can be gained by perplexed, we conducted a case study focused on LLMs for code generation using an additional tool we built to help with the analysis of code models called codetokenizer. Specifically, we explore success and failure cases at the token level of code LLMs under different scenarios pertaining to the type of coding structure the model is predicting, e.g., a variable name or operator, and how predicting of internal verses external method invocations impact performance. From this analysis, we found that our studied code LLMs had their worst performance on coding structures where the code was not syntactically correct. Additionally, we found the models to generally perform worse at predicting internal method invocations than external ones. We have open sourced both of these tools to allow the research community to better understand LLMs in general and LLMs for code generation.

cs.SE↗

arXiv2024

Stable Code Technical Report

We introduce Stable Code, the first in our new-generation of code language models series, which serves as a general-purpose base code language model targeting code completion, reasoning, math, and other software engineering-based tasks. Additionally, we introduce an instruction variant named Stable Code Instruct that allows conversing with the model in a natural chat interface for performing question-answering and instruction-based tasks. In this technical report, we detail the data and training procedure leading to both models. Their weights are available via Hugging Face for anyone to download and use at https://huggingface.co/stabilityai/stable-code-3b and https://huggingface.co/stabilityai/stable-code-instruct-3b. This report contains thorough evaluations of the models, including multilingual programming benchmarks, and the MT benchmark focusing on multi-turn dialogues. At the time of its release, Stable Code is the state-of-the-art open model under 3B parameters and even performs comparably to larger models of sizes 7 billion and 15 billion parameters on the popular Multi-PL benchmark. Stable Code Instruct also exhibits state-of-the-art performance on the MT-Bench coding tasks and on Multi-PL completion compared to other instruction tuned models. Given its appealing small size, we also provide throughput measurements on a number of edge devices. In addition, we open source several quantized checkpoints and provide their performance metrics compared to the original model.

cs.CL↗

arXiv2024

Stable LM 2 1.6B Technical Report

We introduce StableLM 2 1.6B, the first in a new generation of our language model series. In this technical report, we present in detail the data and training procedure leading to the base and instruction-tuned versions of StableLM 2 1.6B. The weights for both models are available via Hugging Face for anyone to download and use. The report contains thorough evaluations of these models, including zero- and few-shot benchmarks, multilingual benchmarks, and the MT benchmark focusing on multi-turn dialogues. At the time of publishing this report, StableLM 2 1.6B was the state-of-the-art open model under 2B parameters by a significant margin. Given its appealing small size, we also provide throughput measurements on a number of edge devices. In addition, we open source several quantized checkpoints and provide their performance metrics compared to the original model.

cs.CL↗

arXiv2023

Intelligent Software Tooling for Improving Software Development

Software has eaten the world with many of the necessities and quality of life services people use requiring software. Therefore, tools that improve the software development experience can have a significant impact on the world such as generating code and test cases, detecting bugs, question and answering, etc., The success of Deep Learning (DL) over the past decade has shown huge advancements in automation across many domains, including Software Development processes. One of the main reasons behind this success is the availability of large datasets such as open-source code available through GitHub or image datasets of mobile Graphical User Interfaces (GUIs) with RICO and ReDRAW to be trained on. Therefore, the central research question my dissertation explores is: In what ways can the software development process be improved through leveraging DL techniques on the vast amounts of unstructured software engineering artifacts?

cs.SE↗

arXiv2023

MASC: A Tool for Mutation-Based Evaluation of Static Crypto-API Misuse Detectors

While software engineers are optimistically adopting crypto-API misuse detectors (or crypto-detectors) in their software development cycles, this momentum must be accompanied by a rigorous understanding of crypto-detectors' effectiveness at finding crypto-API misuses in practice. This demo paper presents the technical details and usage scenarios of our tool, namely Mutation Analysis for evaluating Static Crypto-API misuse detectors (MASC). We developed $12$ generalizable, usage based mutation operators and three mutation scopes, namely Main Scope, Similarity Scope, and Exhaustive Scope, which can be used to expressively instantiate compilable variants of the crypto-API misuse cases. Using MASC, we evaluated nine major crypto-detectors, and discovered $19$ unique, undocumented flaws. We designed MASC to be configurable and user-friendly; a user can configure the parameters to change the nature of generated mutations. Furthermore, MASC comes with both Command Line Interface and Web-based front-end, making it practical for users of different levels of expertise.

cs.CR↗

arXiv2023

Gravitationally-induced entanglement in cold atoms

A promising route to testing quantum gravity in the laboratory is to look for gravitationally-induced entanglement (GIE) between two or more quantum matter systems. Proposals for such tests have principally used microsolid systems, with highly non-classical states, such as N00N states or highly-squeezed states. Here, we consider, for the first time, GIE between two atomic gas interferometers as a test of quantum gravity. We propose placing the two interferometers next to each other in parallel and looking for correlations in the number of atoms at the output ports as evidence of GIE and quantum gravity. GIE is possible without challenging macroscopic superposition states, such as N00N or Schr\"odinger cat states, and instead there can be just classical-like 'coherent' states of atoms. This requires the total mass of the atom interferometers to be on the Planck mass scale, and long integration times. However, with current state-of-the-art quantum squeezing in cold atoms, we argue that the mass scale can be reduced to approachable levels and detail how such a mass scale can be achieved in the near future.

quant-ph↗

arXiv2023

Toward a Theory of Causation for Interpreting Neural Code Models

Neural Language Models of Code, or Neural Code Models (NCMs), are rapidly progressing from research prototypes to commercial developer tools. As such, understanding the capabilities and limitations of such models is becoming critical. However, the abilities of these models are typically measured using automated metrics that often only reveal a portion of their real-world performance. While, in general, the performance of NCMs appears promising, currently much is unknown about how such models arrive at decisions. To this end, this paper introduces $do_{code}$, a post hoc interpretability method specific to NCMs that is capable of explaining model predictions. $do_{code}$ is based upon causal inference to enable programming language-oriented explanations. While the theoretical underpinnings of $do_{code}$ are extensible to exploring different model properties, we provide a concrete instantiation that aims to mitigate the impact of spurious correlations by grounding explanations of model behavior in properties of programming languages. To demonstrate the practical benefit of $do_{code}$, we illustrate the insights that our framework can provide by performing a case study on two popular deep learning architectures and ten NCMs. The results of this case study illustrate that our studied NCMs are sensitive to changes in code syntax. All our NCMs, except for the BERT-like model, statistically learn to predict tokens related to blocks of code (\eg brackets, parenthesis, semicolon) with less confounding bias as compared to other programming language constructs. These insights demonstrate the potential of $do_{code}$ as a useful method to detect and facilitate the elimination of confounding bias in NCMs.

cs.SE↗

arXiv2023

Translating Video Recordings of Complex Mobile App UI Gestures into Replayable Scenarios

Screen recordings of mobile applications are easy to obtain and capture a wealth of information pertinent to software developers (e.g., bugs or feature requests), making them a popular mechanism for crowdsourced app feedback. Thus, these videos are becoming a common artifact that developers must manage. In light of unique mobile development constraints, including swift release cycles and rapidly evolving platforms, automated techniques for analyzing all types of rich software artifacts provide benefit to mobile developers. Unfortunately, automatically analyzing screen recordings presents serious challenges, due to their graphical nature, compared to other types of (textual) artifacts. To address these challenges, this paper introduces V2S+, an automated approach for translating video recordings of Android app usages into replayable scenarios. V2S+ is based primarily on computer vision techniques and adapts recent solutions for object detection and image classification to detect and classify user gestures captured in a video, and convert these into a replayable test scenario. Given that V2S+ takes a computer vision-based approach, it is applicable to both hybrid and native Android applications. We performed an extensive evaluation of V2S+ involving 243 videos depicting 4,028 GUI-based actions collected from users exercising features and reproducing bugs from a collection of over 90 popular native and hybrid Android apps. Our results illustrate that V2S+ can accurately replay scenarios from screen recordings, and is capable of reproducing $\approx$ 90.2% of sequential actions recorded in native application scenarios on physical devices, and $\approx$ 83% of sequential actions recorded in hybrid application scenarios on emulators, both with low overhead. A case study with three industrial partners illustrates the potential usefulness of V2S+ from the viewpoint of developers.

cs.SE↗