Search

A standard assumption in the study of logarithmic structures is "fineness", but this assumption is not preserved by intersections, fiber products, and more general limits. We explain how a coherent logarithmic scheme $X$ has a natural decomposition into "fine faces" -- a collection of fine logarithmic subschemes. This allows for importation of the techniques of tropical geometry into the study of more general logarithmic schemes. We explain the relevance of the construction in enumerative geometry and moduli via a range of examples. Techniques developed in the study of binomial schemes lead to effective algorithms for computing the fine faces., Comment: 24 pages including bibliography. All comments are welcome

Allostery is a fundamental property of proteins that represents the functional coupling between distantly located sites. In different manifestations, this property underlies signal transduction, gene expression, and regulation -- elementary reactions in networks comprising cellular information and metabolic processing systems. In this work, we present a reduced theoretical model for allostery that encompasses the many diverse mechanisms described in past work. What emerges is a basic classification of different forms of allostery into two groups: those that couple distant sites through soft normal modes within a single state, and those that possess multiple states, with effective interactions between sites emerging only upon transitioning between these states. This work serves to unify the extensive past theoretical work on the phenomenology and machinery of allostery in proteins.

We define the logarithmic tautological rings of the moduli spaces of Deligne-Mumford stable curves (together with a set of additive generators lifting the decorated strata classes of the standard tautological rings). While these algebras are infinite dimensional, a connection to polyhedral combinatorics via a new theory of homological piecewise polynomials allows an effective study. A complete calculation is given in genus 0 via the algebra of piecewise polynomials on the cone stack of the associated Artin fan (lifting Keel's presentation of the Chow ring of $\overline{\mathcal{M}}_{0,n}$). Counterexamples to the simplest generalizations in genus 1 are presented. We show, however, that the structure of the log tautological rings is determined by the complete knowledge of all relations in the standard tautological rings of the moduli spaces of curves. In particular, Pixton's conjecture concerning relations in the standard tautological rings lifts to a complete conjecture for relations in the log tautological rings of the moduli spaces of curves. Several open questions are discussed. We develop the entire theory of logarithmic tautological classes in the context of arbitrary smooth normal crossings pairs $(X,D)$ with explicit formulas for intersection products. As a special case, we give an explicit set of additive generators of the full logarithmic Chow ring of $(X,D)$ in terms of Chow classes on the strata of $X$ and piecewise polynomials on the cone stack., Comment: 70 pages. Comments are welcome

In recent years, with the rapid advancement of large language models (LLMs), multi-agent systems have become increasingly more capable of practical application. At the same time, the software development industry has had a number of new AI-powered tools developed that improve the software development lifecycle (SDLC). Academically, much attention has been paid to the role of multi-agent systems to the SDLC. And, while single-agent systems have frequently been examined in real-world applications, we have seen comparatively few real-world examples of publicly available commercial tools working together in a multi-agent system with measurable improvements. In this experiment we test context sharing between Crowdbotics PRD AI, a tool for generating software requirements using AI, and GitHub Copilot, an AI pair-programming tool. By sharing business requirements from PRD AI, we improve the code suggestion capabilities of GitHub Copilot by 13.8% and developer task success rate by 24.5% -- demonstrating a real-world example of commercially-available AI systems working together with improved outcomes., Comment: 7 pages, 8 figures

We investigate the realizability of balanced functions on tropical curves, establishing new sufficient criteria for superabundant functions on genus two curves, analogous to the well-spacedness condition in genus one. We find that realizability is sensitive to the precise locations of conjugate and Weierstrass points on the tropical curve. The key input is a combinatorial comparison of semistable limit theorems for maps of curves. Amini-Baker-Brugall\'e-Rabinoff previously showed that realizability of functions is equivalent to ``modifiability'' to a tropical admissible cover. The resulting criteria are typically inexplicit; we develop combinatorial techniques to derive explicit, verifiable criteria from these. We then develop a dimensional reduction technique to deduce statements about maps to $\mathbb{R}^r$ from ones about maps to $\mathbb{R}$. By proving directly that modifiability and well-spacedness are equivalent in genus one, we obtain a new proof that well-spaced maps are realizable. Along the way, we explain how the modifiability criterion can be viewed as a comparison result for properness statements for moduli of relative maps and admissible covers., Comment: 38 pages, 16 figures

Climate change is a pressing global concern for governments, corporations, and citizens alike. This concern underscores the necessity for these entities to accurately assess the climate impact of manufacturing goods and providing services. Tools like process life cycle analysis (pLCA) are used to evaluate the climate impact of production, use, and disposal, from raw material mining through end-of-life. pLCA further enables practitioners to look deeply into material choices or manufacturing processes for individual parts, sub-assemblies, assemblies, and the final product. Reliable and detailed data on the life cycle stages and processes of the product or service under study are not always available or accessible, resulting in inaccurate assessment of climate impact. To overcome the data limitation and enhance the effectiveness of pLCA to generate an improved environmental impact profile, we are adopting an innovative strategy to identify alternative parts, products, and components that share similarities in terms of their form, function, and performance to serve as qualified substitutes. Focusing on enterprise electronics hardware, we propose a semi-supervised learning-based framework to identify substitute parts that leverages product bill of material (BOM) data and a small amount of component-level qualified substitute data (positive samples) to generate machine knowledge graph (MKG) and learn effective embeddings of the components that constitute electronic hardware. Our methodology is grounded in attributed graph embeddings and introduces a strategy to generate biased negative samples to significantly enhance the training process. We demonstrate improved performance and generalization over existing published models.

This paper presents a radar target tracking framework for addressing main-beam range deception jamming attacks using random finite sets (RFSs). Our system handles false alarms and detections with false range information through multiple hypothesis tracking (MHT) to resolve data association uncertainties. We focus on range gate pull-off (RGPO) attacks, where the attacker adds positive delays to the radar pulse, thereby mimicking the target trajectory while appearing at a larger distance from the radar. The proposed framework incorporates knowledge about the spatial behavior of the attack into the assumed RFS clutter model and uses only position information without relying on additional signal features. We present an adaptive solution that estimates the jammer-induced biases to improve tracking accuracy as well as a simpler non-adaptive version that performs well when accurate priors on the jamming range are available. Furthermore, an expression for RGPO attack detection is derived, where the adaptive solution offers superior performance. The presented strategies provide tracking resilience against multiple RGPO attacks in terms of position estimation accuracy and jamming detection without degrading tracking performance in the absence of jamming., Comment: 12 pages, 7 figures

This paper explores a novel approach to model strategies for flattening wrinkled cloth learning from humans. A human participant study was conducted where the participants were presented with various wrinkle types and tasked with flattening the cloth using the fewest actions possible. A camera and Aruco marker were used to capture images of the cloth and finger movements, respectively. The human strategies for flattening the cloth were modeled using a supervised regression neural network, where the cloth images served as input and the human actions as output. Before training the neural network, a series of image processing techniques were applied, followed by Principal Component Analysis (PCA) to extract relevant features from each image and reduce the input dimensionality. This reduction decreased the model's complexity and computational cost. The actions predicted by the neural network closely matched the actual human actions on an independent data set, demonstrating the effectiveness of neural networks in modeling human actions for flattening wrinkled cloth., Comment: 6 Pages

Brain tumor is a life-threatening problem and hampers the normal functioning of the human body. The average five-year relative survival rate for malignant brain tumors is 35.6 percent. For proper diagnosis and efficient treatment planning, it is necessary to detect the brain tumor in early stages. Due to advancement in medical imaging technology, the brain images are taken in different modalities. The ability to extract relevant characteristics from magnetic resonance imaging (MRI) scans is a crucial step for brain tumor classifiers. Several studies have proposed various strategies to extract relevant features from different modalities of MRI to predict the growth of abnormal tumors. Most techniques used conventional methods of image processing for feature extraction and machine learning for classification. More recently, the use of deep learning algorithms in medical imaging has resulted in significant improvements in the classification and diagnosis of brain tumors. Since tumors are located at different regions of the brain, localizing the tumor and classifying it to a particular category is a challenging task. The objective of this project is to develop a predictive system for brain tumor detection using machine learning(ensembling).

Load instructions often limit instruction-level parallelism (ILP) in modern processors due to data and resource dependences they cause. Prior techniques like Load Value Prediction (LVP) and Memory Renaming (MRN) mitigate load data dependence by predicting the data value of a load instruction. However, they fail to mitigate load resource dependence as the predicted load instruction gets executed nonetheless. Our goal in this work is to improve ILP by mitigating both load data dependence and resource dependence. To this end, we propose a purely-microarchitectural technique called Constable, that safely eliminates the execution of load instructions. Constable dynamically identifies load instructions that have repeatedly fetched the same data from the same load address. We call such loads likely-stable. For every likely-stable load, Constable (1) tracks modifications to its source architectural registers and memory location via lightweight hardware structures, and (2) eliminates the execution of subsequent instances of the load instruction until there is a write to its source register or a store or snoop request to its load address. Our extensive evaluation using a wide variety of 90 workloads shows that Constable improves performance by 5.1% while reducing the core dynamic power consumption by 3.4% on average over a strong baseline system that implements MRN and other dynamic instruction optimizations (e.g., move and zero elimination, constant and branch folding). In presence of 2-way simultaneous multithreading (SMT), Constable's performance improvement increases to 8.8% over the baseline system. When combined with a state-of-the-art load value predictor (EVES), Constable provides an additional 3.7% and 7.8% average performance benefit over the load value predictor alone, in the baseline system without and with 2-way SMT, respectively., Comment: To appear in the proceedings of 51st International Symposium on Computer Architecture (ISCA)

In this work, we present Dyna-5G, a dynamic, self-organizing 5G New Radio (5G-NR) network designed for massive Machine-to-Machine (M2M) networks. Traditional 5G NR networks, characterized by their centralized architecture, face challenges in supporting applications that require dynamic, decentralized communication, such as autonomous vehicles and drone swarms for emergency responses. These scenarios often suffer from the centralized model's single point of failure, undermining the reliability required in critical and fully autonomous applications. Dyna-5G addresses these challenges by allowing each device in the network to function as either part of the Radio Access Network (RAN) and Core Network, or as User Equipment (UE), thus maintaining network functionality even when conventional infrastructure components are compromised. Dyna-5G has built-in mechanisms carefully designed specifically for M2M networks, such as failure-recovery and ad-hoc entry and exit. We demonstrate the performance and feasibility of Dyna-5G using a custom-built testbed that simulates real-world missions, demonstrating our network's robustness, adaptability, and failure recovery capabilities. The results indicate that our entire 5G network model can fully re-organize in 6 seconds at maximum, without compromising the mission.

Software-Defined Networks (SDN) are the standard architecture for network deployment. Intrusion Detection Systems (IDS) are a pivotal part of this technology as networks become more vulnerable to new and sophisticated attacks. Machine Learning (ML)-based IDS are increasingly seen as the most effective approach to handle this issue. However, IDS datasets suffer from high class imbalance, which impacts the performance of standard ML models. We propose Sequential Binary Classification (SBC) - an algorithm for multi-class classification to address this issue. SBC is a hierarchical cascade of base classifiers, each of which can be modelled on any general binary classifier. Extensive experiments are reported on benchmark datasets that evaluate the performance of SBC under different scenarios.

This technical report describes the design and training of novel speculative decoding draft models, for accelerating the inference speeds of large language models in a production environment. By conditioning draft predictions on both context vectors and sampled tokens, we can train our speculators to efficiently predict high-quality n-grams, which the base model then accepts or rejects. This allows us to effectively predict multiple tokens per inference forward pass, accelerating wall-clock inference speeds of highly optimized base model implementations by a factor of 2-3x. We explore these initial results and describe next steps for further improvements., Comment: Original upload 4/29/24, updated 6/6/24 with additional references to concurrent work

Conventional approaches to enhancing movement coordination, such as providing instructions and visual feedback, are often inadequate in complex motor tasks with multiple degrees of freedom (DoFs). To effectively address coordination deficits in such complex motor systems, it becomes imperative to develop interventions grounded in a model of human motor learning; however, modeling such learning processes is challenging due to the large DoFs. In this paper, we present a computational motor learning model that leverages the concept of motor synergies to extract low-dimensional learning representations in the high-dimensional motor space and the internal model theory of motor control to capture both fast and slow motor learning processes. We establish the model's convergence properties and validate it using data from a target capture game played by human participants. We study the influence of model parameters on several motor learning trade-offs such as speed-accuracy, exploration-exploitation, satisficing, and flexibility-performance, and show that the human motor learning system tunes these parameters to optimize learning and various output performance metrics., Comment: 22 pages (single column), 9 figures

We study stable maps to normal crossings pairs with possibly negative tangency orders. There are two independent models: punctured Gromov-Witten theory of pairs and orbifold Gromov-Witten theory of root stacks with extremal ages. Exploiting the tropical structure of the punctured mapping space, we define and study a new virtual class for the punctured theory. This arises as a refined intersection product on the Artin fan, and produces a distinguished sector of punctured Gromov-Witten invariants. Restricting to genus zero, we show that these invariants coincide with the orbifold invariants, first for smooth pairs, and then for normal crossings pairs after passing to a sufficiently refined blowup. This builds on previous work to provide a complete picture of the logarithmic-orbifold comparison in genus zero, which is compatible with splitting and thus allows for the wholesale importation of orbifold techniques, including boundary recursion and torus localisation. Contemporaneous work of Johnston uses the comparison to give a new proof of the associativity of the Gross-Siebert intrinsic mirror ring., Comment: 50 pages. Comments are welcome. v2: minor changes

We propose a logarithmic enhancement of the Gromov-Witten/Donaldson-Thomas correspondence, with descendants, and study the behaviour of the correspondence under simple normal crossings degenerations. The formulation of the logarithmic correspondence requires a matching of tangency conditions and relative insertions. This is achieved via a version of the Nakajima basis for the cohomology of the Hilbert schemes of points on logarithmic surfaces. We then establish a strong form of the degeneration formula in logarithmic DT theory - the numerical DT invariants of the general fiber of a degeneration are determined by the numerical DT invariants attached to strata of the special fiber. The GW version of this result, which we also prove here, is a strengthening of the currently known formulas. A key role is played by a certain exotic class of insertions, introduced here, and can be thought of as non-local incidence conditions coupled across multiple boundary strata of the target geometry. Finally, we prove compatiblity of the new logarithmic GW/DT correspondence with degenerations, and in particular, knowledge of the conjecture on the strata of the special fiber of a degeneration implies it on the general fiber. Several examples are included to illustrate the nature and utility of the formula., Comment: 89 pages. Comments are welcome