arXiv subjects

Zhenkun Li

Publications and source records attributed to Zhenkun Li.

At least 19 recordsLinked to original sources

arXiv2026

SU(2) representations and a large surgery formula

A knot $K\subset S^3$ is called $SU(2)$-abundant if it satisfies two conditions: first, for all but finitely many $r\in\mathbb{Q}\backslash\{0\}$, there exists an irreducible representation $π_1(S^3_r(K))\to SU(2)$; second, any slope $r=u/v\neq 0$ for which $S^3_r(K)$ admits no irreducible $SU(2)$ representation must satisfy $Δ_K(ζ^2)= 0$ for some $u$-th root of unity $ζ$. We show that if a nontrivial knot $K\subset S^3$ is not $SU(2)$-abundant then it is a prime knot whose Alexander polynomial $Δ_K(t)$ has coefficients restricted to $\{-1,0,1\}$. This implies, in particular, that all hyperbolic alternating knots are $SU(2)$-abundant. Our proof hinges on a large surgery formula connecting instanton knot homology $KHI(S^3,K)$ and framed instanton homology $I^\sharp(S^3_n(K))$ for integers $n$ satisfying $|n|\ge 2g(K)+1$. Using this technique, we derive several interesting results in instanton Floer homology: for any Berge knot $K$, the spaces $KHI(S^3,K)$ and $\widehat{HFK}(S^3,K)$ have identical dimension; for any dual knot $K_r\subset S^3_r(K)$ of a Berge knot $K$ with $r> 2g(K)-1$, we prove $\dim_\mathbb{C}KHI(S^3_r(K),K_r)=|H_1(S^3_r(K);\mathbb{Z})|$; and for any genus-one alternating knot $K$ and any $r\in\mathbb{Q}\backslash\{0\}$, the spaces $I^\sharp(S^3_r(K))$ and $\widehat{HF}(S_r^3(K))$ have equal dimension.

math.GT↗

arXiv2026

2-torsion in instanton Floer homology

This paper studies the existence of $2$-torsion in instanton Floer homology with $\mathbb{Z}$ coefficients for closed $3$-manifolds and singular knots. First, we show that the non-existence of $2$-torsion in the framed instanton Floer homology $I^\sharp(S_n^3(K);\mathbb{Z})$ of any nonzero integral $n$-surgery along a knot $K$ in $S^3$ would imply that $K$ is fibered. Also, we show that $I^\sharp(S_{r}^3(K);\mathbb{Z})$ for any nontrivial $K$ with $r=1,1/2,1/4$ always has $2$-torsion. These two results indicate that the existence of $2$-torsion is expected to be a generic phenomenon for Dehn surgeries along knots. Second, we show that for genus-one knots with nontrivial Alexander polynomials and for unknotting-number-one knots, the unreduced singular instanton knot homology $I^\sharp(S^3,K;\mathbb{Z})$ always has $2$-torsion. Finally, some crucial lemmas that help us demonstrate the existence of $2$-torsion are motivated by analogous results in Heegaard Floer theory, which may be of independent interest. In particular, we show that, for a knot $K$ in $S^3$, if there is a nonzero rational number $r$ such that the dual knot $\widetilde{K}_r$ inside $S^3_r(K)$ is Floer simple, then $S^3_r(K)$ must be an L-space and $K$ must be an L-space knot.

math.GT↗

arXiv2026

Guts of nearly fibered knots

The guts of a knot is an invariant defined for the knot complement by Agol-Zhang. Nearly fibered knots, which are defined as knots whose Floer homology has dimension two in the top Alexander grading, were introduced by Baldwin-Sivek. In this note, we provide three models for the guts of nearly fibered knots in the $3$-sphere. As a corollary, the nearly fibered condition can be purely topologically characterized and is independent of the specific version of Floer theory.

math.GT↗

arXiv2025

Instanton 2-torsion and fibered knots

We prove that the unreduced singular instanton homology $I^\sharp(Y,K;\mathbb{Z})$ has $2$-torsion for any null-homologous fibered knot $K$ of genus $g>0$ in a closed $3$-manifold $Y$ except for $\#^{2g}S^1\times S^2$. The main technical result is a formula of $I^\sharp(Y,K;\mathbb{C})$ via sutured instanton theory, by which we can compare the dimensions of $I^\sharp(Y,K;\mathbb{F}_2)$ and $I^\sharp(Y,K;\mathbb{C})$. As a byproduct, we show that $I^\sharp(S^3,K;\mathbb{C})$ for a knot $K\subset S^3$ admitting lens space surgeries is determined by the Alexander polynomial, while some special cases of torus knots have been previously studied by many people. Another byproduct is that the next-to-top Alexander grading summand of instanton knot homology $KHI(S^3,K,g(K)-1)$ is non-vanishing when $K$ has unknotting number one, which generalizes the Baldwin--Sivek's result in the fibered case. Finally, we discuss the relation to the Heegaard Floer theory.

math.GT↗

arXiv2025

Knot surgery formulae for instanton Floer homology II: applications

This is a companion paper to earlier work of the authors, which proved an integral surgery formula for framed instanton homology. First, we present an enhancement of the large surgery formula, a rational surgery formula for null-homologous knots in any 3-manifold, and a formula encoding a large portion of $I^\sharp(S^3_0(K))$. Second, we use the integral surgery formula to study the framed instanton homology of many 3-manifolds: Seifert fibered spaces with nonzero orbifold degrees, especially nontrivial circle bundles over any orientable surface, surgeries on a family of alternating knots and all twisted Whitehead doubles, and splicings with twist knots. Finally, we use the previous techniques and computations to study almost L-space knots, ${\it i.e.}$, the knots $K\subset S^3$ with $\dim I^\sharp(S_n^3(K))=n+2$ for some $n\in\mathbb{N}_+$. We show that an almost L-space knot of genus at least $2$ is fibered and strongly quasi-positive, and a genus-one almost L-space knot must be either the figure eight or the mirror of the $5_2$ knot in Rolfsen's knot table.

math.GT↗

arXiv2025

Instanton dimensions of knot surgeries over arbitrary fields

Suppose $K \subset S^3$ is a knot and suppose $p$ and $q$ are co-prime integers with $q\ge 1$. For any field $\mathbb{K}$, we establish a dimension formula for the framed instanton homology of knot surgeries: $$ \dim I^\sharp(S^3_{p/q}(K); \mathbb{K}) = q \cdot r_{\mathbb{K}}(K) + |p - q \cdot ν^\sharp_{\mathbb{K}}(K)| $$ for certain integers $r_{\mathbb{K}}(K)$ and $ν^\sharp_{\mathbb{K}}(K)$, except possibly when $p/q = ν^\sharp_{\mathbb{K}}(K)$ and $ν^\sharp_{\mathbb{K}}(K)$ is even. This formula generalizes the result of Baldwin--Sivek from the case $\mathbb{K} = \mathbb{C}$ to arbitrary fields. Based on the result for $\mathbb{K} = \mathbb{Z}/2$, we obtain that $S^3_{p/q}(K)$ is not $SU(2)$-abelian for any knot $K$ other than the unknot and the right-handed trefoil whenever $p/q \in [0,6)$ and $p \in \{ a^e, 2a^e \}$ for some prime number $a$ and natural number $e$, thereby extending existing results for $p/q \in [0,5]$ and $p = a^e$. A byproduct of the techniques developed in this paper is that we generalize the distance-two surgery exact triangle by Culler--Daemi--Xie and Daemi--Miller-Eismeier--Lidman from $\mathbb{Z}/2$ coefficients to any coefficient ring.

math.GT↗

arXiv2025

Can LLM Agents Really Debate? A Controlled Study of Multi-Agent Debate in Logical Reasoning

Multi-agent debate (MAD) has recently emerged as a promising framework for improving the reasoning performance of large language models (LLMs). Yet, whether LLM agents can genuinely engage in deliberative reasoning, beyond simple ensembling or majority voting, remains unclear. We address this question through a controlled study using the Knight--Knave--Spy logic puzzle, which enables precise, step-wise evaluation of debate outcomes and processes under verifiable ground truth. We systematically set up six structural and cognitive factors, including agent team size, composition, confidence visibility, debate order, debate depth, and task difficulty, to disentangle their respective effects on collective reasoning. Our results show that intrinsic reasoning strength and group diversity are the dominant drivers of debate success, while structural parameters such as order or confidence visibility offer limited gains. Beyond outcomes, process-level analyses identify key behavioral patterns: majority pressure suppresses independent correction, effective teams overturn incorrect consensus, and rational, validity-aligned reasoning most strongly predicts improvement. These findings provide valuable insights into how and why LLM debates succeed or fail, offering guidance for designing interpretable and truth-seeking multi-agent reasoning systems.

cs.MA↗

arXiv2025

PartnerMAS: An LLM Hierarchical Multi-Agent Framework for Business Partner Selection on High-Dimensional Features

High-dimensional decision-making tasks, such as business partner selection, involve evaluating large candidate pools with heterogeneous numerical, categorical, and textual features. While large language models (LLMs) offer strong in-context reasoning capabilities, single-agent or debate-style systems often struggle with scalability and consistency in such settings. We propose PartnerMAS, a hierarchical multi-agent framework that decomposes evaluation into three layers: a Planner Agent that designs strategies, Specialized Agents that perform role-specific assessments, and a Supervisor Agent that integrates their outputs. To support systematic evaluation, we also introduce a curated benchmark dataset of venture capital co-investments, featuring diverse firm attributes and ground-truth syndicates. Across 140 cases, PartnerMAS consistently outperforms single-agent and debate-based multi-agent baselines, achieving up to 10--15\% higher match rates. Analysis of agent reasoning shows that planners are most responsive to domain-informed prompts, specialists produce complementary feature coverage, and supervisors play an important role in aggregation. Our findings demonstrate that structured collaboration among LLM agents can generate more robust outcomes than scaling individual models, highlighting PartnerMAS as a promising framework for high-dimensional decision-making in data-rich domains.

cs.MA↗

arXiv2025

Surgeries on knots and tight contact structures

For any knot $K$ in $S^3$ and any positive rational $r$, we show that smooth $(-r)$-surgery on $K$ always admits a tight contact structure. More specifically, the tightness is detected by the non-vanishing Heegaard Floer contact invariant.

math.GT↗

arXiv2025

$SU(2)$-representations of Branched Covers

We study the existence of irreducible $SU(2)$-representations for cyclic branched covers of knots in $S^3$. Our main result establishes that if $K$ is a non-trivial prime knot and $d$ is an integer such that $d \geq 2$ and $Σ_d(K)$ is an integer homology sphere, then $π_1(Σ_d(K))$ admits an irreducible $SU(2)$-representation, whenever $K$ satisfies one of two conditions: either $K$ is $2$-periodic, or $K$ can be represented as the closure of a tangle adapted to a $d\times d$ SICUP matrix. The first condition leverages a commuting trick for covering spaces to realize higher-degree branched covers as 2-fold covers, allowing us to apply recent results of Kronheimer-Mrowka and others. The second condition uses equivariant surgery descriptions and the $ν^\sharp$ invariant from instanton Floer homology. As applications, we provide new infinite families of hyperbolic integer homology spheres admitting irreducible representations, including examples where previously known criteria fail.

math.GT↗

arXiv2025

Instanton 2-torsion and Dehn surgeries

In our earlier work on $2$-torsion in instanton Floer homology, we considered only integral surgeries on a knot $K\subset S^3$ and showed that the absence of $2$-torsion forces $K$ to be fibered. The present paper extends the result to all rational surgeries. We prove that if the framed instanton homology $I^{\sharp}(S^3_r(K);\mathbb{Z})$ is $2$-torsion-free for some $r\in \mathbb{Q}_+$, then $K$ is an instanton L-space knot and $r>2g(K)-1$. Leveraging this $2$-torsion perspective, we also obtain new small-surgery obstructions: If either $S^{3}_{5}(K)$ or $S^{3}_{11/2}(K)$ is $SU(2)$-abelian, then $K$ must be the unknot or the right-handed trefoil. This result sharpens the small-$SU(2)$-abelian surgery theorems of Kronheimer--Mrowka, Baldwin--Sivek, and Baldwin--Li--Sivek--Ye.

math.GT↗

arXiv2025

Know the Ropes: A Heuristic Strategy for LLM-based Multi-Agent System Design

Single-agent LLMs hit hard limits--finite context, role overload, and brittle domain transfer. Conventional multi-agent fixes soften those edges yet expose fresh pains: ill-posed decompositions, fuzzy contracts, and verification overhead that blunts the gains. We therefore present Know-The-Ropes (KtR), a framework that converts domain priors into an algorithmic blueprint hierarchy, in which tasks are recursively split into typed, controller-mediated subtasks, each solved zero-shot or with the lightest viable boost (e.g., chain-of-thought, micro-tune, self-check). Grounded in the No-Free-Lunch theorem, KtR trades the chase for a universal prompt for disciplined decomposition. On the Knapsack problem (3-8 items), three GPT-4o-mini agents raise accuracy from 3% zero-shot to 95% on size-5 instances after patching a single bottleneck agent. On the tougher Task-Assignment problem (6-15 jobs), a six-agent o3-mini blueprint hits 100% up to size 10 and 84% on sizes 13-15, versus 11% zero-shot. Algorithm-aware decomposition plus targeted augmentation thus turns modest models into reliable collaborators--no ever-larger monoliths required.

cs.AI↗

arXiv2024

On Legendrian representatives of non-fibered knots

We show that in $(S^3,ξ_{std})$ if $K$ is a non-trivial knot that realizes the three-dimensional Thurston-Bennequin bound (i.e. $K$ has a Legendrian representative $Λ$ with $tb(Λ)-rot(Λ)=2g(K)-1$), then $K$ has a Legendrian representative $L$ with $tb=0$. Moreover, this result can be easily generalized to contact manifolds that uniquely represent the associated contact invariants. This is the first result on Legendrian representatives of non-fibered knots in $3-$manifolds other than $S^3$. We also show that if $K$ is a nearly fibered knot in $S^3$ then $τ(K)=g(K)$ implies that $K$ realizes the three-dimensional Thurston-Bennequin bound.

math.GT↗

arXiv2024

Knot surgery formulae for instanton Floer homology I: the main theorem

We prove an integral surgery formula for framed instanton homology $I^\sharp(Y_m(K))$ for any knot $K$ in a $3$-manifold $Y$ with $[K]=0\in H_1(Y;\mathbb{Q})$ and $m\neq 0$. Though the statement is similar to Ozsváth-Szabó's integral surgery formula for Heegaard Floer homology, the proof is new and based on sutured instanton homology $SHI$ and the octahedral lemma in the derived category. As a corollary, we obtain an exact triangle between $I^\sharp(Y_m(K))$, $I^\sharp(Y_{m+k}(K))$ and $k$ copies of $I^\sharp(Y)$ for any $m\neq 0$ and large $k$. In the proof of the formula, we discover many new exact triangles for sutured instanton homology and relate some surgery cobordism map to the sum of bypass maps, which are of independent interest. In a companion paper, we derive many applications and computations based on the integral surgery formula.

math.GT↗

arXiv2024

An enhanced Euler characteristic of sutured instanton homology

For a balanced sutured manifold $(M,γ)$, we construct a decomposition of $SHI(M,γ)$ with respect to torsions in $H=H_1(M;\mathbb{Z})$, which generalizes the decomposition of $I^\sharp(Y)$ in previous work of the authors. This decomposition can be regarded as a candidate for the counterpart of the torsion spin$^c$ decompositions in $SFH(M,γ)$. Based on this decomposition, we define an enhanced Euler characteristic $χ_{\rm en}(SHI(M,γ))\in\mathbb{Z}[H]/\pm H$ and prove that $χ_{\rm en}(SHI(M,γ))=χ(SFH(M,γ))$. This provides a better lower bound on $\dim_\mathbb{C}SHI(M,γ)$ than the graded Euler characteristic $χ_{\rm gr}(SHI(M,γ))$. As applications, we prove instanton knot homology detects the unknot in any instanton L-space and show that the conjecture $KHI(Y,K)\cong \widehat{HFK}(Y,K)$ holds for all $(1,1)$-L-space knots and constrained knots in lens spaces, which include all torus knots and many hyperbolic knots in lens spaces.

math.GT↗

arXiv2024

Instanton Floer homology, sutures, and Euler characteristics

This is a companion paper to an earlier work of the authors. In this paper, we provide an axiomatic definition of Floer homology for balanced sutured manifolds and prove that the graded Euler characteristic $χ_{\rm gr}$ of this homology is fully determined by the axioms we proposed. As a result, we conclude that $χ_{\rm gr}(SHI(M,γ))=χ_{\rm gr}(SFH(M,γ))$ for any balanced sutured manifold $(M,γ)$. In particular, for any link $L$ in $S^3$, the Euler characteristic $χ_{\rm gr}(KHI(S^3,L))$ recovers the multi-variable Alexander polynomial of $L$, which generalizes the knot case. Combined with the authors' earlier work, we provide more examples of $(1,1)$-knots in lens spaces whose $KHI$ and $\widehat{HFK}$ have the same dimension. Moreover, for a rationally null-homologous knot in a closed oriented 3-manifold $Y$, we construct canonical $\mathbb{Z}_2$-gradings on $KHI(Y,K)$, the decomposition of $I^\sharp(Y)$ discussed in the previous paper, and the minus version of instanton knot homology $\underline{\rm KHI}^-(Y,K)$ introduced by the first author.

math.GT↗

arXiv2023

Knot cobordism, torsion order and framed instanton homology

We construct cobordism maps for the \textit{minus} version of instanton knot homology associated to a \textit{specially decorated} knot cobordisms of arbitrary genus between two null-homologous knots in closed oriented $3$-manifolds. As an application of our construction, we recover an inequality between the torsion order of knots in instanton theory, which was originally established in Heegaard Floer theory by work of Juhász, Miller, and Zemke. We further use this inequality to compute the framed instanton Floer homology of any non-zero Dehn surgeries along an alternating knot of bridge index at most $3$.

math.GT↗

arXiv2023

Some Computations on Instanton Knot Homology

In a recent paper, the first author and his collaborator developed a method to compute an upper bound of the dimension of instanton Floer homology via Heegaard Diagrams of 3-manifolds. For a knot inside S3, we further develop an algorithm that can compute an upper bound of the dimension of instanton knot homology from knot diagrams. We test the effectiveness of the algorithm and found that for all knots up to seven crossings, the algorithm provides sharp bounds. In the second half of the paper, we show that, if the instanton knot Floer homology of a knot has a specified form, then the knot must an instanton L-space knot.

math.GT↗