arXiv subjects

Jared Coleman

Publications and source records attributed to Jared Coleman.

18 recordsLinked to original sources

arXiv2026

Search and Rescue on the Plane

We study a planar variant of the search and rescue problem whereby an agent starting at an arbitrary position $P_{\theta,r} = (r\cos\theta, r\sin\theta)$ in the plane must locate an object at an unknown position on the positive $x$-axis and deliver it to the origin. Our main contribution is to characterize the optimal form of any competitive algorithm, derive closed-form expressions for the competitive ratio, and identify a critical angle $\theta^* \approx 15.6^\circ$ which yields a phase transition to optimal competitive search and delivery in the following sense. For each angle $-\pi \leq \theta \leq \pi$ we compute a checkpoint (landing position on the $x$-axis) where the agent must go first prior to initiating a search on the $x$-axis in order to optimize the competitive ratio of search and delivery. We show that if $|\theta| \geq \theta^*$ then the checkpoint is at the origin, while if $|\theta| < \theta^*$ then the agent should land at the checkpoint $(r \cdot k_{|\theta|}, 0)$ on the $x$-axis, where $k_{|\theta|}$ is a real number given by an explicit formula we present.

cs.DM↗

arXiv2026

ncsim: A Lightweight Simulator for Networked Edge Computing with Wireless Interference Modeling

Evaluating directed acyclic graph (DAG) task schedulers for wireless edge computing requires jointly modeling compute placement and wireless interference, yet many evaluation setups simplify the wireless coupling. This can lead to rank inversions: a scheduler selected under an interference-free model need not remain best when concurrent wireless traffic is included. We present ncsim, a lightweight discrete-event simulator that combines DAG workflow scheduling with an analytical IEEE 802.11 contention and hidden-interference model in a single Python package. A 108-run factorial experiment on grid networks and DAG workflows finds rank inversions in 6 of 18 scenarios (33.3%). Solo retains single-link MAC overhead, while Full adds concurrent-link wireless effects. Selecting a policy under Solo and executing it under Full incurs a maximum scheduling regret (makespan penalty relative to the best evaluated policy under Full) of 52.3%. Validation includes Bianchi's published numerical and graphical results and fixed-rate ns-3 experiments. The default model closely tracks homogeneous contention. The scheduler results show how wireless coupling can change policy selection within the stated flow model.

cs.DC↗

arXiv2026

Studying the Effect of Schedule Preemption on Dynamic Task Graph Scheduling

Dynamic scheduling of task graphs is often addressed without revisiting prior task allocations, with a primary focus on minimizing makespan. We study controlled schedule preemption, introducing the Last-K Preemption model, which selectively reschedules recent task graphs while preserving earlier allocations. Using synthetic, RIoTBench, WFCommons, and adversarial workloads, we compare preemptive, non-preemptive, and partial-preemptive strategies across makespan, fairness, utilization, and runtime. Results show moderate preemption can match most makespan and utilization gains of full preemption while maintaining fairness and low overhead.

cs.DC↗

arXiv2025

Crowd-SFT: Crowdsourcing for LLM Alignment

Large Language Models (LLMs) increasingly rely on Supervised Fine-Tuning (SFT) and Reinforcement Learning from Human Feedback (RLHF) to align model responses with human preferences. While RLHF employs a reinforcement learning approach with a separate reward model, SFT uses human-curated datasets for supervised learning. Both approaches traditionally depend on small, vetted groups of annotators, making them costly, prone to bias, and limited in scalability. We propose an open, crowd-sourced fine-tuning framework that addresses these limitations by enabling broader feedback collection for SFT without extensive annotator training. Our framework promotes incentive fairness via a point-based reward system correlated with Shapley values and guides model convergence through iterative model updates. Our multi-model selection framework demonstrates up to a 55% reduction in target distance over single-model selection, enabling subsequent experiments that validate our point-based reward mechanism's close alignment with Shapley values (a well-established method for attributing individual contributions) thereby supporting fair and scalable participation.

cs.HC↗

arXiv2025

Linear Search with Probabilistic Detection and Variable Speeds

We present results on new variants of the famous linear search (or cow-path) problem that involves an agent searching for a target with unknown position on the infinite line. We consider the variant where the agent can move either at speed $1$ or at a slower speed $v \in [0, 1)$. When traveling at the slower speed $v$, the agent is guaranteed to detect the target upon passing through its location. When traveling at speed $1$, however, the agent, upon passing through the target's location, detects it with probability $p \in [0, 1]$. We present algorithms and provide upper bounds for the competitive ratios for three cases separately: when $p=0$, $v=0$, and when $p,v \in (0,1)$. We also prove that the provided algorithm for the $p=0$ case is optimal.

cs.DM↗

arXiv2025

Multimodal Search on a Line

Inspired by the diverse set of technologies used in underground object detection and imaging, we introduce a novel multimodal linear search problem whereby a single searcher starts at the origin and must find a target that can only be detected when the searcher moves through its location using the correct of $p$ possible search modes. The target's location, its distance $d$ from the origin, and the correct search mode are all initially unknown to the searcher. We prove tight upper and lower bounds on the competitive ratio for this problem. Specifically, we show that when $p$ is odd, the optimal competitive ratio is given by $2p+3+\sqrt{8(p+1)}$, whereas when $p$ is even, the optimal competitive ratio is given by $c$: the unique solution to $(c-1)^4-4p(c+1)^2(c-p-1)=0$ in the interval $\left[2p+1+\sqrt{8p},\infty\right)$. This solution $c$ has the explicit bounds $2p+3+\sqrt{8(p-1)}\leq c\leq 2p+3+\sqrt{8p}$. The optimal algorithms we propose require the searcher to move infinitesimal distances and change directions infinitely many times within finite intervals. To better suit practical applications, we also propose an approximation algorithm with a competitive ratio of $c+\varepsilon$ (where $c$ is the optimal competitive ratio and $\varepsilon > 0$ is an arbitrarily small constant). This algorithm involves the searcher moving finite distances and changing directions a finite number of times within any finite interval.

cs.DM↗

arXiv2024

Workflows Community Summit 2024: Future Trends and Challenges in Scientific Workflows

The Workflows Community Summit gathered 111 participants from 18 countries to discuss emerging trends and challenges in scientific workflows, focusing on six key areas: time-sensitive workflows, AI-HPC convergence, multi-facility workflows, heterogeneous HPC environments, user experience, and FAIR computational workflows. The integration of AI and exascale computing has revolutionized scientific workflows, enabling higher-fidelity models and complex, time-sensitive processes, while introducing challenges in managing heterogeneous environments and multi-facility data dependencies. The rise of large language models is driving computational demands to zettaflop scales, necessitating modular, adaptable systems and cloud-service models to optimize resource utilization and ensure reproducibility. Multi-facility workflows present challenges in data movement, curation, and overcoming institutional silos, while diverse hardware architectures require integrating workflow considerations into early system design and developing standardized resource management tools. The summit emphasized improving user experience in workflow systems and ensuring FAIR workflows to enhance collaboration and accelerate scientific discovery. Key recommendations include developing standardized metrics for time-sensitive workflows, creating frameworks for cloud-HPC integration, implementing distributed-by-design workflow modeling, establishing multi-facility authentication protocols, and accelerating AI integration in HPC workflow management. The summit also called for comprehensive workflow benchmarks, workflow-specific UX principles, and a FAIR workflow maturity model, highlighting the need for continued collaboration in addressing the complex challenges posed by the convergence of AI, HPC, and multi-facility research environments.

cs.DC↗

arXiv2024

LLM-Assisted Rule Based Machine Translation for Low/No-Resource Languages

We propose a new paradigm for machine translation that is particularly useful for no-resource languages (those without any publicly available bilingual or monolingual corpora): LLM-RBMT (LLM-Assisted Rule Based Machine Translation). Using the LLM-RBMT paradigm, we design the first language education/revitalization-oriented machine translator for Owens Valley Paiute (OVP), a critically endangered Indigenous American language for which there is virtually no publicly available data. We present a detailed evaluation of the translator's components: a rule-based sentence builder, an OVP to English translator, and an English to OVP translator. We also discuss the potential of the paradigm, its limitations, and the many avenues for future research that it opens up.

cs.CL↗

arXiv2024

Optimal Delivery with a Faulty Drone

We introduce and study a new cooperative delivery problem inspired by drone-assisted package delivery. We consider a scenario where a drone, en route to deliver a package to a destination (a point on the plane), unexpectedly loses communication with its central command station. The command station cannot know whether the drone's system has wholly malfunctioned or merely experienced a communications failure. Consequently, a second, helper drone must be deployed to retrieve the package to ensure successful delivery. The central question of this study is to find the optimal trajectory for this second drone. We demonstrate that the optimal solution relies heavily on the relative spatial positioning of the command station, the destination point, and the last known location of the disconnected drone.

cs.DM↗

arXiv2024

Linear Search for an Escaping Target with Unknown Speed

We consider linear search for an escaping target whose speed and initial position are unknown to the searcher. A searcher (an autonomous mobile agent) is initially placed at the origin of the real line and can move with maximum speed $1$ in either direction along the line. An oblivious mobile target that is moving away from the origin with an unknown constant speed $v<1$ is initially placed by an adversary on the infinite line at distance $d$ from the origin in an unknown direction. We consider two cases, depending on whether $d$ is known or unknown. The main contribution of this paper is to prove a new lower bound and give algorithms leading to new upper bounds for search in these settings. This results in an optimal (up to lower order terms in the exponent) competitive ratio in the case where $d$ is known and improved upper and lower bounds for the case where $d$ is unknown. Our results solve an open problem proposed in [Coleman et al., Proc. OPODIS 2022].

cs.DM↗

arXiv2024

Parameterized Task Graph Scheduling Algorithm for Comparing Algorithmic Components

Scheduling distributed applications modeled as directed, acyclic task graphs to run on heterogeneous compute networks is a fundamental (NP-Hard) problem in distributed computing for which many heuristic algorithms have been proposed over the past decades. Many of these algorithms fall under the list-scheduling paradigm, whereby the algorithm first computes priorities for the tasks and then schedules them greedily to the compute node that minimizes some cost function. Thus, many algorithms differ from each other only in a few key components (e.g., the way they prioritize tasks, their cost functions, where the algorithms consider inserting tasks into a partially complete schedule, etc.). In this paper, we propose a generalized parametric list-scheduling algorithm that allows mixing and matching different algorithmic components to produce 72 unique algorithms. We benchmark these algorithms on four datasets to study the individual and combined effects of different algorithmic components on performance and runtime.

cs.DC↗

arXiv2024

PISA: An Adversarial Approach To Comparing Task Graph Scheduling Algorithms

Scheduling a task graph representing an application over a heterogeneous network of computers is a fundamental problem in distributed computing. It is known to be not only NP-hard but also not polynomial-time approximable within a constant factor. As a result, many heuristic algorithms have been proposed over the past few decades. Yet it remains largely unclear how these algorithms compare to each other in terms of the quality of schedules they produce. We identify gaps in the traditional benchmarking approach to comparing task scheduling algorithms and propose a simulated annealing-based adversarial analysis approach called PISA to help address them. We also introduce SAGA, a new open-source library for comparing task scheduling algorithms. We use SAGA to benchmark 15 algorithms on 16 datasets and PISA to compare the algorithms in a pairwise manner. Algorithms that appear to perform similarly on benchmarking datasets are shown to perform very differently on adversarially chosen problem instances. Interestingly, the results indicate that this is true even when the adversarial search is constrained to selecting among well-structured, application-specific problem instances. This work represents an important step towards a more general understanding of the performance boundaries between task scheduling algorithms on different families of problem instances.

cs.DC↗

arXiv2023

DARSAN: A Decentralized Review System Suitable for NFT Marketplaces

We introduce DARSAN, a decentralized review system designed for Non-Fungible Token (NFT) marketplaces, to address the challenge of verifying the quality of highly resalable products with few verified buyers by incentivizing unbiased reviews. DARSAN works by iteratively selecting a group of reviewers (called ``experts'') who are likely to both accurately predict the objective popularity and assess some subjective quality of the assets uniquely associated with NFTs. The system consists of a two-phased review process: a ``pre-listing'' phase where only experts can review the product, and a ``pre-sale'' phase where any reviewer on the system can review the product. Upon completion of the sale, DARSAN distributes incentives to the participants and selects the next generation of experts based on the performance of both experts and non-expert reviewers. We evaluate DARSAN through simulation and show that, once bootstrapped with an initial set of appropriately chosen experts, DARSAN favors honest reviewers and improves the quality of the expert pool over time without any external intervention even in the presence of potentially malicious participants.

cs.CY↗

arXiv2022

Line Search for an Oblivious Moving Target

Consider search on an infinite line involving an autonomous robot starting at the origin of the line and an oblivious moving target at initial distance $d \geq 1$ from it. The robot can change direction and move anywhere on the line with constant maximum speed $1$ while the target is also moving on the line with constant speed $v>0$ but is unable to change its speed or direction. The goal is for the robot to catch up to the target in as little time as possible. The classic case where $v=0$ and the target's initial distance $d$ is unknown to the robot is the well-studied ``cow-path problem''. Alpert and Gal gave an optimal algorithm for the case where a target with unknown initial distance $d$ is moving away from the robot with a known speed $v<1$. In this paper we design and analyze search algorithms for the remaining possible knowledge situations, namely, when $d$ and $v$ are known, when $v$ is known but $d$ is unknown, when $d$ is known but $v$ is unknown, and when both $v$ and $d$ are unknown. Furthermore, for each of these knowledge models we consider separately the case where the target is moving away from the origin and the case where it is moving toward the origin. We design algorithms and analyze competitive ratios for all eight cases above. The resulting competitive ratios are shown to be optimal when the target is moving towards the origin as well as when $v$ is known and the target is moving away from the origin.

cs.DC↗

arXiv2022

Delivery to Safety with Two Cooperating Robots

Two cooperating, autonomous mobile robots with arbitrary nonzero max speeds are placed at arbitrary initial positions in the plane. A remotely detonated bomb is discovered at some source location and must be moved to a safe distance away from its initial location as quickly as possible. In the Bomb Squad problem, the robots cooperate by communicating face-to-face in order to pick up the bomb from the source and carry it away to the boundary of a disk centered at the source in the shortest possible time. The goal is to specify trajectories which define the robots' paths from start to finish and their meeting points which enable face-to-face collaboration by exchanging information and passing the bomb from robot to robot. We design algorithms reflecting the robots' knowledge about orientation and each other's speed and location. In the offline case, we design an optimal algorithm. For the limited knowledge cases, we provide online algorithms which consider robots' level of agreement on orientation as per OneAxis and NoAxis models, and knowledge of the boundary as per Visible, Discoverable, and Invisible. In all cases, we provide upper and lower bounds for the competitive ratios of the online problems.

cs.DC↗

arXiv2021

Message Delivery in the Plane by Robots with Different Speeds

We study a fundamental cooperative message-delivery problem on the plane. Assume $n$ robots which can move in any direction, are placed arbitrarily on the plane. Robots each have their own maximum speed and can communicate with each other face-to-face (i.e., when they are at the same location at the same time). There are also two designated points on the plane, $S$ (the source) and $D$ (the destination). The robots are required to transmit the message from the source to the destination as quickly as possible by face-to-face message passing. We consider both the offline setting where all information (the locations and maximum speeds of the robots) are known in advance and the online setting where each robot knows only its own position and speed along with the positions of $S$ and $D$. In the offline case, we discover an important connection between the problem for two-robot systems and the well-known Apollonius circle which we employ to design an optimal algorithm. We also propose a $\sqrt 2$ approximation algorithm for systems with any number of robots. In the online setting, we provide an algorithm with competitive ratio $\frac 17 \left( 5+ 4 \sqrt{2} \right)$ for two-robot systems and show that the same algorithm has a competitive ratio less than $2$ for systems with any number of robots. We also show these results are tight for the given algorithm. Finally, we give two lower bounds (employing different arguments) on the competitive ratio of any online algorithm, one of $1.0391$ and the other of $1.0405$.

cs.DC↗

arXiv2021

The Pony Express Communication Problem

We introduce a new problem which we call the Pony Express problem. n robots with differing speeds are situated over some domain. A message is placed at some commonly known point. Robots can acquire the message either by visiting its initial position, or by encountering another robot that has already acquired it. The robots must collaborate to deliver the message to a given destination. The objective is to deliver the message in minimum time. In this paper we study the Pony Express problem on the line where n robots are arbitrarily deployed along a finite segment. The robots have different speeds and can move in both directions. We are interested in both offline centralized and online distributed algorithms. In the online case, we assume the robots have limited knowledge of the initial configuration. In particular, the robots do not know the initial positions and speeds of the other robots nor even their own position and speed. They do, however, know the direction on the line in which to find the message and have the ability to compare speeds when they meet. First, we study the Pony Express problem where the message is initially placed at one endpoint of a segment and must be delivered to the other endpoint. We provide an O(n log n) running time offline algorithm as well as an optimal online algorithm. Then we study the Half-Broadcast problem where the message is at the center and must be delivered to either one of the endpoints of the segment [-1,1]. We provide an offline algorithm running in O(n^2 log n) time and we provide an online algorithm that attains a competitive ratio of 3/2 which we show is the best possible. Finally, we study the Broadcast problem where the message is at the center and must be delivered to both endpoints of the segment [-1,1]. Here we give an FPTAS in the offline case and an online algorithm that attains a competitive ratio of 9/5, which we show is tight.

cs.DC↗

arXiv2020

Minimizing The Maximum Distance Traveled To Form Patterns With Systems of Mobile Robots

In the pattern formation problem, robots in a system must self-coordinate to form a given pattern, regardless of translation, rotation, uniform-scaling, and/or reflection. In other words, a valid final configuration of the system is a formation that is \textit{similar} to the desired pattern. While there has been no shortage of research in the pattern formation problem under a variety of assumptions, models, and contexts, we consider the additional constraint that the maximum distance traveled among all robots in the system is minimum. Existing work in pattern formation and closely related problems are typically application-specific or not concerned with optimality (but rather feasibility). We show the necessary conditions any optimal solution must satisfy and present a solution for systems of three robots. Our work also led to an interesting result that has applications beyond pattern formation. Namely, a metric for comparing two triangles where a distance of $0$ indicates the triangles are similar, and $1$ indicates they are \emph{fully dissimilar}.

cs.CG↗