SearcharxivSearch

arXiv subjects

Henry Liu

Publications and source records attributed to Henry Liu.

At least 19 recordsLinked to original sources

Wall-crossing for equivariant DT4 invariants

We prove the wall-crossing formula conjectured by Gross--Joyce--Tanaka for equivariant enumerative invariants of CY4 categories equipped with framing functors. We also establish a version for stable pairs with fixed-determinant obstruction theories, as used in earlier applications by the first-named author. The main technical ingredient, which we develop in this work, is a construction of well-behaved CY4 pullback virtual classes using Jouanolou devices.

math.AG

On degree powers in the degenerate Tur\'an problem

Given a graph $G$ with degree sequence $d_{1},\ldots,d_{n}$ and a positive real number $p$, let $e_{p}(G)=\sum_{i=1}^{n} d_{i}^{p}$. For a fixed family of graphs $\mathcal F$, let $ex_{p}(n, \mathcal F)$ denote the maximum value of $e_{p}(G)$ over all $\mathcal F$-free graphs $G$ on $n$ vertices. In 2000, Caro and Yuster introduced the following Tur\'an-type problem: For a positive integer $p$ and a fixed graph $F$, determine $ex_{p}(n, F)$, and characterize the extremal graphs $G$ on $n$ vertices that attain $ex_p(n, F)$. Recently, Gao, Liu, Ma and Pikhurko proved that $ex_{p}(n, \mathcal F)=(\tau(\mathcal F)-1+o(1))n^p$ for real $p>\frac{1}{1-\alpha}$, where $\mathcal F$ is a degenerate family of graphs with classical Tur\'an number $ex(n, \mathcal F)=O(n^{1+\alpha})$ for some $\alpha\in[0,1)$, and $\tau(\mathcal F)$ is the minimum size of an independent vertex cover over all bipartite graphs $F\in\mathcal F$. Based on their method, we obtain a stability result for $ex_{p}(n, \mathcal F)$, and prove that all extremal graphs must contain the complete bipartite graph $K_{\tau(\mathcal F)-1,n-\tau(\mathcal F)+1}$ when $n$ is sufficiently large. Our results can be used to deduce all previously known results about $ex_{p}(n, F)$ when $F$ is a bipartite graph and $n$ is sufficiently large. We also obtain several new exact results for $ex_{p}(n, F)$, namely, when $F$ is an even cycle, a complete bipartite graph, a discrete hypercube, a caterpillar forest, and a spider forest.

math.CO

Learning as Reasoning Unfolds: Progressive Rollout Allocation for Efficient Reinforcement Learning

Reinforcement learning with verifiable rewards (RLVR) has emerged as a highly effective framework for improving LLM reasoning, with methods such as GRPO among its most successful instantiations. However, GRPO relies on repeated generation of long chain-of-thought rollouts. Training time scales with the number of rollouts, a large fraction of which are uninformative. Thus, GRPO is computationally expensive and unstable. To mitigate this, existing approaches either generate a larger pool of rollouts and filter the most informative prompts, or leverage historical signals for filtering at later stages of training. These strategies offer modest performance gains, but slow down the overall process. To address this, we propose VarIance Guided Online Rollout allocation (VIGOR) which instead of allocating a fixed rollout budget per example, begins with a small number of rollouts for all examples in a batch and iteratively allocates additional rollouts to those with the highest group reward variance until a fixed total rollout budget is reached. Theoretically, we show that under RLVR, reward variance controls the gradient magnitude, and derive VIGOR's closed-form speedup ratio over GRPO, which grows with refinement rounds under Pareto-distributed reward variance. Experiments on mathematical reasoning and coding tasks show that VIGOR reaches target accuracy with up to 2.3$\times$ fewer rollouts on math, reaches GRPO's final coding full pass rate with 1.49$\times$ fewer rollouts, and improves the coding average test pass rate by 3.4 points.

cs.AI

Functional Safety Analysis for Infrastructure-Enabled Depot Autonomy System

This paper presents the functional safety analysis for an Infrastructure-Enabled Depot Autonomy (IX-DA) system. The IX-DA system automates the marshalling of delivery vehicles within a controlled depot environment, navigating connected autonomous vehicles (CAVs) between drop-off zones, service stations (washing, calibration, charging, loading), and pick-up zones without human intervention. We describe the system architecture comprising three principal subsystems -- the connected autonomous vehicle, the infrastructure sensing and compute layer, and the human operator interface -- and derive their functional requirements. Using ISO 26262-compliant Hazard Analysis and Risk Assessment (HARA) methodology, we identify eight hazardous events, evaluate them across different operating scenarios, and assign Automotive Safety Integrity Levels~(ASILs) ranging from Quality Management (QM) to ASIL C. Six safety goals are derived and allocated to vehicle and infrastructure subsystems. The analysis demonstrates that high-speed uncontrolled operation imposes the most demanding safety requirements (ASIL C), while controlled low-speed operation reduces most goals to QM, offering a practical pathway for phased deployment.

cs.ET

Wall-crossing for invariants of equivariant 3CY categories

We provide a wall-crossing framework for operational enumerative invariants of equivariant 3-Calabi--Yau categories arising from virtual cycles. The strategy follows ideas of Joyce's ``universal'' wall-crossing framework arXiv:2111.04694, using the authors' symmetrized pullback technique to preserve the symmetry of the (almost-perfect) obstruction theories throughout. As an application, we define and study wall-crossings of simple type between operational equivariant Donaldson--Thomas (DT), Pandharipande--Thomas (PT), and Bryan--Steinberg (BS) vertices. In particular, we give an explicit DT/PT descendent vertex correspondence in the Calabi--Yau limit. As another application, we construct and prove wall-crossing formulas for operational refined semistable Vafa--Witten invariants.

math.AG

IMProofBench: Benchmarking AI on Research-Level Mathematical Proof Generation

As the mathematical capabilities of large language models (LLMs) improve, it becomes increasingly important to evaluate their performance on research-level tasks at the frontier of mathematical knowledge. However, existing benchmarks are limited, as they focus solely on final-answer questions or high-school competition problems. To address this gap, we introduce IMProofBench, a private benchmark consisting of 77 peer-reviewed problems developed by expert mathematicians. Each problem requires a detailed proof and is paired with subproblems that have final answers, supporting both an evaluation by human experts and a large-scale quantitative analysis through automated grading. Furthermore, unlike prior benchmarks, the evaluation setup simulates a realistic research environment: models operate in an agentic framework with tools like web search for literature review and mathematical software such as SageMath. Our results show that current LLMs can already solve a significant percentage of research-level questions. IMProofBench will continue to evolve as a dynamic benchmark in collaboration with the mathematical community, ensuring its relevance for evaluating the next generation of LLMs.

cs.CL

List rainbow connection number of graphs

An edge-coloured path is rainbow if all of its edges have distinct colours. Let $G$ be a connected graph. The rainbow connection number of $G$, denoted by $rc(G)$, is the minimum number of colours in an edge-colouring of $G$ such that, any two vertices are connected by a rainbow path. The strong rainbow connection number of $G$, denoted by $src(G)$, is the minimum number of colours in an edge-colouring of $G$ such that, any two vertices are connected by a rainbow geodesic (i.e., a path of shortest length). These two notions of connectivity of graphs were introduced by Chartrand, Johns, McKeon and Zhang in 2008. In this paper, we introduce the list rainbow connection number $rc^\ell(G)$, and the list strong rainbow connection number $src^\ell(G)$. These two parameters are the versions of $rc(G)$ and $src(G)$ that involve list edge-colourings. Among our results, we will determine the list rainbow connection number and list strong rainbow connection number of some specific graphs. We will also characterise all pairs of positive integers $a$ and $b$ such that, there exists a connected graph $G$ with $src(G)=a$ and $src^\ell(G)=b$, and similarly for the pair $rc^\ell$ and $src^\ell$. Finally, we propose the question of whether or not we have $rc(G)=rc^\ell(G)$, for all connected graphs $G$.

math.CO

OMEGA: A Low-Latency GNN Serving System for Large Graphs

Graph Neural Networks (GNNs) have been widely adopted for their ability to compute expressive node representations in graph datasets. However, serving GNNs on large graphs is challenging due to the high communication, computation, and memory overheads of constructing and executing computation graphs, which represent information flow across large neighborhoods. Existing approximation techniques in training can mitigate the overheads but, in serving, still lead to high latency and/or accuracy loss. To this end, we propose OMEGA, a system that enables low-latency GNN serving for large graphs with minimal accuracy loss through two key ideas. First, OMEGA employs selective recomputation of precomputed embeddings, which allows for reusing precomputed computation subgraphs while selectively recomputing a small fraction to minimize accuracy loss. Second, we develop computation graph parallelism, which reduces communication overhead by parallelizing the creation and execution of computation graphs across machines. Our evaluation with large graph datasets and GNN models shows that OMEGA significantly outperforms state-of-the-art techniques.

cs.DC

Hitchin systems and their quantization

This is an expanded version of the notes by the second author of the lectures on Hitchin systems and their quantization given by the first author at the Beijing Summer Workshop in Mathematics and Mathematical Physics ``Integrable Systems and Algebraic Geometry" (BIMSA-2024).

math.AG

Mass-Balance MRV for Carbon Dioxide Removal by Enhanced Rock Weathering: Methods, Simulation, and Inference

Carbon dioxide will likely need to be removed from the atmosphere to avoid significant future warming and climate change. Technologies are being developed to remove large quantities of carbon from the atmosphere. Enhanced rock weathering (ERW), where fine-grained silicate minerals are spread on soil, is a promising carbon removal method that can also support crop yields and maintain overall soil health. Quantifying the amount of carbon removed by ERW is crucial for understanding the potential of ERW globally and for building trust in commercial operations. However, reliable and scalable quantification in complex media like soil is challenging and there is not yet a consensus on the best method of doing so. Here we discuss mass-balance methods, where stocks of base cations in soil are monitored over time to infer the amount of inorganic carbon brought into solution by weathering reactions. First, we review the fundamental concepts of mass-balance methods and explain different ways of approaching the mass-balance problem. Then we discuss experimental planning and data collection, suggesting some best practices. Next, we present a software package designed to facilitate a range of tasks in ERW like uncertainty analysis, planning field trials, and validating statistical methods. Finally, we briefly review ways of estimating carbon removal using mass balance before discussing some advantages of Bayesian inference in this context and presenting an example Bayesian model. The model is fit to simulated data and recovers the correct answer with a clear representation of uncertainty.

stat.AP

Rainbow cycles through specified vertices

An edge-coloured cycle is rainbow if the edges have distinct colours. Let $G$ be a graph such that any $k$ vertices lie in a cycle of $G$. The $k$-rainbow cycle index of $G$, denoted by $crx_k(G)$, is the minimum number of colours required to colour the edges of $G$ such that, for every set $S$ of $k$ vertices in $G$, there exists a rainbow cycle in $G$ containing $S$. In this paper, we will first prove some results about the parameter $crx_k(G)$ for general graphs $G$. One of the results is a classification of all graphs $G$ such that $crx_k(G)=e(G)$, for $k=1,2$. We will also determine $crx_k(G)$ for some specific graphs $G$, including wheels, complete graphs, complete bipartite and multipartite graphs, and discrete cubes.

math.CO

Invariance of elliptic genus under wall-crossing

Wall-crossing formulas for various flavors of elliptic genus can be obtained using master spaces. We give a topological criterion which implies that such wall-crossing formulas are trivial. Applications are given for: GIT quotients, following Thaddeus; moduli of sheaves, following Mochizuki; Donaldson-Thomas and Vafa-Witten theory, following Joyce and Tanaka-Thomas respectively.

math.AG

Monochromatic $k$-connection of graphs

An edge-coloured path is monochromatic if all of its edges have the same colour. For a $k$-connected graph $G$, the monochromatic $k$-connection number of $G$, denoted by $mc_k(G)$, is the maximum number of colours in an edge-colouring of $G$ such that, any two vertices are connected by $k$ internally vertex-disjoint monochromatic paths. In this paper, we shall study the parameter $mc_k(G)$. We obtain bounds for $mc_k(G)$, for general graphs $G$. We also compute $mc_k(G)$ exactly when $k$ is small, and $G$ is a graph on $n$ vertices, with a spanning $k$-connected subgraph having the minimum possible number of edges, namely $\lceil\frac{kn}{2}\rceil$. We prove a similar result when $G$ is a bipartite graph.

math.CO

The 3-fold K-theoretic DT/PT vertex correspondence holds

We prove the 3-fold DT/PT correspondence for K-theoretic vertices via wall-crossing techniques. We provide two different setups, following Mochizuki and following Joyce; both reduce the problem to q-combinatorial identities on word rearrangements. An important technical step is the construction of symmetric almost-perfect obstruction theories (APOTs) on auxiliary moduli stacks, e.g. master spaces, from the symmetric DT or PT obstruction theory. For this, we introduce symmetrized pullbacks of symmetric obstruction theories along smooth morphisms of Artin stacks.

math.AG

Semistable refined Vafa-Witten invariants

For any smooth complex projective surface $S$, we construct semistable refined Vafa-Witten invariants of $S$ which prove the main conjecture of arXiv:1810.00078. This is done by extending part of Joyce's universal wall-crossing formalism to equivariant K-theory, and to moduli stacks with symmetric obstruction theories, particularly moduli stacks of sheaves on Calabi-Yau threefolds. An important technical tool which we introduce is the symmetrized pullback, along smooth morphisms, of symmetric obstruction theories.

math.AG

The 4-fold Pandharipande--Thomas vertex

We give a conjectural but full and explicit description of the (K-theoretic) equivariant vertex for Pandharipande--Thomas stable pairs on toric Calabi--Yau 4-folds, by identifying torus-fixed loci as certain quiver Grassmannians and prescribing a canonical half of the tangent-obstruction theory. For any number of non-trivial legs, the DT/PT vertex correspondence can then be verified by computer in low degrees.

math.AG

Multiplicative vertex algebras and quantum loop algebras

We define a multiplicative version of vertex coalgebras and show that various equivariant K-theoretic constructions of Hall algebras (KHAs) also admit a compatible multiplicative vertex coalgebra structure. In particular, this is true of the preprojective KHA of Varagnolo--Vasserot, which is (conjecturally) isomorphic to certain quantum loop algebras.

math.AG

Equivariant K-theoretic enumerative invariants and wall-crossing formulae in abelian categories

We provide a general framework for wall-crossing of equivariant K-theoretic enumerative invariants of appropriate moduli stacks $\mathfrak{M}$, by lifting Joyce's homological universal wall-crossing arXiv:2111.04694 to K-theory and to include equivariance. The primary new tool is that the operational K-homology of $\mathfrak{M}$ is an equivariant multiplicative vertex algebra.

math.AG