arXiv subjects

Jonathan Roberts

Publications and source records attributed to Jonathan Roberts.

At least 19 recordsLinked to original sources

arXiv2026

How Long Is a Piece of String? A Brief Empirical Analysis of Tokenizers

Frontier LLMs are increasingly utilised across academia, society and industry. A commonly used unit for comparing models, their inputs and outputs, and estimating inference pricing is the token. In general, tokens are used as a stable currency, assumed to be broadly consistent across tokenizers and contexts, enabling direct comparisons. However, tokenization varies significantly across models and domains of text, making naive interpretation of token counts problematic. We quantify this variation by providing a comprehensive empirical analysis of tokenization, exploring the compression of sequences to tokens across different distributions of textual data. Our analysis challenges commonly held heuristics about token lengths, finding them to be overly simplistic. We hope the insights of our study add clarity and intuition toward tokenization in contemporary LLMs.

cs.CL↗

arXiv2025

New Orbital Constraints for YSES 1 b and HR 2562 B from High-Precision Astrometry and Planetary Radial Velocities

We present new VLTI/GRAVITY astrometry and updated orbit fits for the directly imaged companions YSES 1 b and HR 2562 B, substellar objects straddling the planet-brown dwarf boundary. Using high-precision astrometry, radial velocity (RV) data, and proper motions, we derive revised orbital parameters with orbitize! arXiv:1910.01756. For YSES 1 b, the inclusion of GRAVITY astrometry and a relative radial velocity measurement from arXiv:2409.16660 overcomes the traditional challenge of constraining eccentricities for distant companions, enabling the first orbit fit and yielding a constrained eccentricity of 0.44 (0.20). This represents the first full orbit fit for the system. Additionally, we calculate a median line-of-sight stellar obliquity of 12 (+11, -8) degrees, providing further insight into the system's dynamical architecture. For HR 2562 B, our analysis agrees with arXiv:2302.04893, confirming a low-eccentricity orbit (0.34 (0.20)) and an inclination of 87 (1) degrees. We find HR 2562 B's orbit to be nearly coplanar with the debris disk, with a mutual inclination of 3.7 (0.3) degrees. For both YSES 1 b and HR 2562 B the lower eccentricities favor an in situ formation scenario over extreme scattering or cloud fragmentation.

astro-ph.EP↗

arXiv2025

ZeroBench: An Impossible Visual Benchmark for Contemporary Large Multimodal Models

Large Multimodal Models (LMMs) exhibit shortfalls when interpreting images and, by some measures, have poorer spatial cognition than young children or animals. Despite this, they attain high scores on many popular visual benchmarks, with headroom rapidly eroded by model progress. This creates a need for difficult benchmarks that remain relevant for longer. We introduce ZeroBench - a lightweight visual reasoning benchmark curated using adversarial filtering to be "impossible" for frontier LMMs at its original release, with initial SotA scores of 0% pass@1 and pass^5. We track progress on ZeroBench over the subsequent year, observing SotA reaching 6% pass^5 and 19% pass@5, indicating the potential longevity of the benchmark. We evaluate 46 LMMs on ZeroBench, compare performance to a human baseline, analyse strengths and weaknesses, chart a year of progress in visual capabilities, and publicly release ZeroBench at https://zerobench.github.io.

cs.CV↗

arXiv2024

GAMEBoT: Transparent Assessment of LLM Reasoning in Games

Large Language Models (LLMs) are increasingly deployed in real-world applications that demand complex reasoning. To track progress, robust benchmarks are required to evaluate their capabilities beyond superficial pattern recognition. However, current LLM reasoning benchmarks often face challenges such as insufficient interpretability, performance saturation or data contamination. To address these challenges, we introduce GAMEBoT, a gaming arena designed for rigorous and transparent assessment of LLM reasoning capabilities. GAMEBoT decomposes complex reasoning in games into predefined modular subproblems. This decomposition allows us to design a suite of Chain-of-Thought (CoT) prompts that leverage domain knowledge to guide LLMs in addressing these subproblems before action selection. Furthermore, we develop a suite of rule-based algorithms to generate ground truth for these subproblems, enabling rigorous validation of the LLMs' intermediate reasoning steps. This approach facilitates evaluation of both the quality of final actions and the accuracy of the underlying reasoning process. GAMEBoT also naturally alleviates the risk of data contamination through dynamic games and head-to-head LLM competitions. We benchmark 17 prominent LLMs across eight games, encompassing various strategic abilities and game characteristics. Our results suggest that GAMEBoT presents a significant challenge, even when LLMs are provided with detailed CoT prompts. Project page: https://visual-ai.github.io/gamebot

cs.CL↗

arXiv2024

Needle Threading: Can LLMs Follow Threads through Near-Million-Scale Haystacks?

As the context limits of Large Language Models (LLMs) increase, the range of possible applications and downstream functions broadens. In many real-world tasks, decisions depend on details scattered across collections of often disparate documents containing mostly irrelevant information. Long-context LLMs appear well-suited to this form of complex information retrieval and reasoning, which has traditionally proven costly and time-consuming. However, although the development of longer context models has seen rapid gains in recent years, our understanding of how effectively LLMs use their context has not kept pace. To address this, we conduct a set of retrieval experiments designed to evaluate the capabilities of 17 leading LLMs, such as their ability to follow threads of information through the context window. Strikingly, we find that many models are remarkably threadsafe: capable of simultaneously following multiple threads without significant loss in performance. Still, for many models, we find the effective context limit is significantly shorter than the supported context length, with accuracy decreasing as the context window grows. Our study also highlights the important point that token counts from different tokenizers should not be directly compared -- they often correspond to substantially different numbers of written characters. We release our code and long-context experimental data.

cs.CL↗

arXiv2024

GRAB: A Challenging GRaph Analysis Benchmark for Large Multimodal Models

Large multimodal models (LMMs) have exhibited proficiencies across many visual tasks. Although numerous well-known benchmarks exist to evaluate model performance, they increasingly have insufficient headroom. As such, there is a pressing need for a new generation of benchmarks challenging enough for the next generation of LMMs. One area that LMMs show potential is graph analysis, specifically, the tasks an analyst might typically perform when interpreting figures such as estimating the mean, intercepts or correlations of functions and data series. In this work, we introduce GRAB, a graph analysis benchmark, fit for current and future frontier LMMs. Our benchmark is predominantly synthetic, ensuring high-quality, noise-free questions. GRAB is comprised of 3284 questions, covering five tasks and 23 graph properties. We evaluate 20 LMMs on GRAB, finding it to be a challenging benchmark, with the highest performing model attaining a score of just 21.0%. Finally, we conduct various ablations to investigate where the models succeed and struggle. We release GRAB and a lightweight GRAB-Lite to encourage progress in this important, growing domain.

cs.CV↗

arXiv2024

Visual Storytelling: A Methodological Approach to Designing and Implementing a Visualisation Poster

We present a design study of developing a visualisation poster. Posters can be difficult to create, and the story on a poster is not always clear. Using a case-study approach we propose three important aspects: the poster should have a clear focus (especially a hero visualisation), envisioning its use helps to drive the important aspects, and third the essence (its fundamental concept and guiding idea) must be clear. We will use case studies that have focused on the use of the Five Design-Sheet method (FdS) as a way to sketch and plan a visualisation, before successfully implementing and creating the visual poster. The case studies serve as a practical illustration of the workflow, offering a means to explain the three key processes involved: (1) comprehending the data, (2) employing a design study with the FdS (Five Design-Sheet), (3) crafting, evaluating and refining the visualisation.

cs.HC↗

arXiv2024

SciFIBench: Benchmarking Large Multimodal Models for Scientific Figure Interpretation

Large multimodal models (LMMs) have proven flexible and generalisable across many tasks and fields. Although they have strong potential to aid scientific research, their capabilities in this domain are not well characterised. A key aspect of scientific research is the ability to understand and interpret figures, which serve as a rich, compressed source of complex information. In this work, we present SciFIBench, a scientific figure interpretation benchmark consisting of 2000 questions split between two tasks across 8 categories. The questions are curated from arXiv paper figures and captions, using adversarial filtering to find hard negatives and human verification for quality control. We evaluate 28 LMMs on SciFIBench, finding it to be a challenging benchmark. Finally, we investigate the alignment and reasoning faithfulness of the LMMs on augmented question sets from our benchmark. We release SciFIBench to encourage progress in this domain.

cs.CV↗

arXiv2023

Charting New Territories: Exploring the Geographic and Geospatial Capabilities of Multimodal LLMs

Multimodal large language models (MLLMs) have shown remarkable capabilities across a broad range of tasks but their knowledge and abilities in the geographic and geospatial domains are yet to be explored, despite potential wide-ranging benefits to navigation, environmental research, urban development, and disaster response. We conduct a series of experiments exploring various vision capabilities of MLLMs within these domains, particularly focusing on the frontier model GPT-4V, and benchmark its performance against open-source counterparts. Our methodology involves challenging these models with a small-scale geographic benchmark consisting of a suite of visual tasks, testing their abilities across a spectrum of complexity. The analysis uncovers not only where such models excel, including instances where they outperform humans, but also where they falter, providing a balanced view of their capabilities in the geographic domain. To enable the comparison and evaluation of future models, our benchmark will be publicly released.

cs.CV↗

arXiv2023

GPT4GEO: How a Language Model Sees the World's Geography

Large language models (LLMs) have shown remarkable capabilities across a broad range of tasks involving question answering and the generation of coherent text and code. Comprehensively understanding the strengths and weaknesses of LLMs is beneficial for safety, downstream applications and improving performance. In this work, we investigate the degree to which GPT-4 has acquired factual geographic knowledge and is capable of using this knowledge for interpretative reasoning, which is especially important for applications that involve geographic data, such as geospatial analysis, supply chain management, and disaster response. To this end, we design and conduct a series of diverse experiments, starting from factual tasks such as location, distance and elevation estimation to more complex questions such as generating country outlines and travel networks, route finding under constraints and supply chain analysis. We provide a broad characterisation of what GPT-4 (without plugins or Internet access) knows about the world, highlighting both potentially surprising capabilities but also limitations.

cs.CL↗

arXiv2023

SATIN: A Multi-Task Metadataset for Classifying Satellite Imagery using Vision-Language Models

Interpreting remote sensing imagery enables numerous downstream applications ranging from land-use planning to deforestation monitoring. Robustly classifying this data is challenging due to the Earth's geographic diversity. While many distinct satellite and aerial image classification datasets exist, there is yet to be a benchmark curated that suitably covers this diversity. In this work, we introduce SATellite ImageNet (SATIN), a metadataset curated from 27 existing remotely sensed datasets, and comprehensively evaluate the zero-shot transfer classification capabilities of a broad range of vision-language (VL) models on SATIN. We find SATIN to be a challenging benchmark-the strongest method we evaluate achieves a classification accuracy of 52.0%. We provide a $\href{https://satinbenchmark.github.io}{\text{public leaderboard}}$ to guide and track the progress of VL models in this important domain.

cs.CV↗

arXiv2021

3D Semantic Mapping from Arthroscopy using Out-of-distribution Pose and Depth and In-distribution Segmentation Training

Minimally invasive surgery (MIS) has many documented advantages, but the surgeon's limited visual contact with the scene can be problematic. Hence, systems that can help surgeons navigate, such as a method that can produce a 3D semantic map, can compensate for the limitation above. In theory, we can borrow 3D semantic mapping techniques developed for robotics, but this requires finding solutions to the following challenges in MIS: 1) semantic segmentation, 2) depth estimation, and 3) pose estimation. In this paper, we propose the first 3D semantic mapping system from knee arthroscopy that solves the three challenges above. Using out-of-distribution non-human datasets, where pose could be labeled, we jointly train depth+pose estimators using selfsupervised and supervised losses. Using an in-distribution human knee dataset, we train a fully-supervised semantic segmentation system to label arthroscopic image pixels into femur, ACL, and meniscus. Taking testing images from human knees, we combine the results from these two systems to automatically create 3D semantic maps of the human knee. The result of this work opens the pathway to the generation of intraoperative 3D semantic mapping, registration with pre-operative data, and robotic-assisted arthroscopy

cs.RO↗

arXiv2021

Arthroscopic Multi-Spectral Scene Segmentation Using Deep Learning

Knee arthroscopy is a minimally invasive surgical (MIS) procedure which is performed to treat knee-joint ailment. Lack of visual information of the surgical site obtained from miniaturized cameras make this surgical procedure more complex. Knee cavity is a very confined space; therefore, surgical scenes are captured at close proximity. Insignificant context of knee atlas often makes them unrecognizable as a consequence unintentional tissue damage often occurred and shows a long learning curve to train new surgeons. Automatic context awareness through labeling of the surgical site can be an alternative to mitigate these drawbacks. However, from the previous studies, it is confirmed that the surgical site exhibits several limitations, among others, lack of discriminative contextual information such as texture and features which drastically limits this vision task. Additionally, poor imaging conditions and lack of accurate ground-truth labels are also limiting the accuracy. To mitigate these limitations of knee arthroscopy, in this work we proposed a scene segmentation method that successfully segments multi structures.

eess.IV↗

arXiv2020

IRIS: A Low Duty Cycle Cross-Layer Protocol for Long-Range Wireless Sensor Networks with Low Power Budget

This paper presents a cross-layer protocol (IRIS) designed for long-range pipeline Wireless Sensor Networks with extremely low power budget, typically seen in a range of monitoring applications. IRIS uses ping packets initiated by a base station to travel through the multi-hop network and carry monitoring information. The protocol is able to operate with less than 1% duty cycle, thereby conforming to ISM band spectrum regulations in the 868MHz band. The duty cycle can be flexibly configured to meet other regulations/power budgets as well as to improve the route forming performance. Simulation results show guaranteed route formation in different network topologies with various protocol configurations. System robustness against unreliable wireless connections and node failures are also demonstrated by simulations.

cs.NI↗

arXiv2020

Composition and Configuration Patterns in Multiple-View Visualizations

Multiple-view visualization (MV) is a layout design technique often employed to help users see a large number of data attributes and values in a single cohesive representation. Because of its generalizability, the MV design has been widely adopted by the visualization community to help users examine and interact with large, complex, and high-dimensional data. However, although ubiquitous, there has been little work to categorize and analyze MVs in order to better understand its design space. As a result, there has been little to no guideline in how to use the MV design effectively. In this paper, we present an in-depth study of how MVs are designed in practice. We focus on two fundamental measures of multiple-view patterns: composition, which quantifies what view types and how many are there; and configuration, which characterizes spatial arrangement of view layouts in the display space. We build a new dataset containing 360 images of MVs collected from IEEE VIS, EuroVis, and PacificVis publications 2011 to 2019, and make fine-grained annotations of view types and layouts for these visualization images. From this data we conduct composition and configuration analyses using quantitative metrics of term frequency and layout topology. We identify common practices around MVs, including relationship of view types, popular view layouts, and correlation between view types and layouts. We combine the findings into a MV recommendation system, providing interactive tools to explore the design space, and support example-based design.

cs.HC↗

arXiv2020

Powderday: Dust Radiative Transfer for Galaxy Simulations

We present Powderday, a flexible, fast, open-source dust radiative transfer package designed to interface with galaxy formation simulations. Powderday builds on FSPS population synthesis models, Hyperion dust radiative transfer, and employs yt to interface between different software packages. We include our stellar population synthesis modeling on the fly, which allows for significant run-time flexibility in the assumed stellar physics. We include a model for nebular line emission that can employ either precomputed Cloudy lookup tables (for efficiency), or direct photoionization calculations for all young stars (for flexibility). The dust content follows either observationally-motivated prescriptions, direct modeling from galaxy formation simulations, or a novel approach that includes the dust content via learning-based algorithms from the SIMBA cosmological galaxy formation simulation. AGN can additionally be included via a range of prescriptions. The output of these models are broadband SEDs, as well as filter-convolved images. Powderday is designed to eliminate last-mile efforts by researchers that employ different hydrodynamic galaxy formation models, and seamlessly interfaces with GIZMO, AREPO, GASOLINE, CHANGA, and ENZO. We demonstrate the capabilities of the code via three applications: a model for the star formation rate (SFR) - infrared luminosity relation in galaxies (including the impact of AGN); the impact of circumstellar dust around AGB stars on the mid-infrared emission from galaxy SEDs; and the impact of galaxy inclination angle on dust attenuation laws.

astro-ph.GA↗

arXiv2019

Real-time Joint Motion Analysis and Instrument Tracking for Robot-Assisted Orthopaedic Surgery

Robotic-assisted orthopaedic surgeries demand accurate, automated leg manipulation for improved spatial accuracy to reduce iatrogenic damage. In this study, we propose novel rigid body designs and an optical tracking volume setup for tracking of the femur, tibia and surgical instruments. Anatomical points inside the leg are measured using Computed Tomography with an accuracy of 0.3mm. Combined with kinematic modelling, we can express these points relative to any frame and across joints to sub-millimetre accuracy. It enables the setup of vectors on the mechanical axes of the femur and tibia for kinematic analysis. Cadaveric experiments are used to verify the tracking of internal anatomies and joint motion analysis. The proposed integrated solution is a first step in the automation of leg manipulation and can be used as a ground-truth for future robot-assisted orthopaedic research.

cs.RO↗

arXiv2019

Optimal Dexterity for a Snake-like Surgical Manipulator using Patient-specific Task-space Constraints in a Computational Design Algorithm

Tendon-driven snake-like arms have been used to create highly dexterous continuum robots so that they can bend around anatomical obstacles to access clinical targets. In this paper, we propose a design algorithm for developing patient-specific surgical continuum manipulators optimized for oriental dexterity constrained by task-space obstacles. The algorithm uses a sampling-based approach to finding the dexterity distribution in the workspace discretized by voxels. The oriental dexterity measured in the region of interest in the task-space formed a fitness function to be optimized through differential evolution. This was implemented in the design of a tendon-driven manipulator for knee arthroscopy. The results showed a feasible design that achieves significantly better dexterity than a rigid tool. This highlights the potential of the proposed method to be used in the process of designing dexterous surgical manipulators in the field.

cs.RO↗