arXiv subjects

Jiahao Qiu

Publications and source records attributed to Jiahao Qiu.

At least 19 recordsLinked to original sources

arXiv2026

Harnessing the Reasoning Economy: A Survey of Efficient Reasoning for Large Language Models

Recent advancements in Large Language Models (LLMs) have significantly enhanced their ability to perform complex reasoning tasks, transitioning from fast and intuitive thinking (System 1) to slow and deep reasoning (System 2). While System 2 reasoning improves task accuracy, it often incurs substantial computational costs due to its slow thinking nature and inefficient or unnecessary reasoning behaviors. In contrast, System 1 reasoning is computationally efficient but leads to suboptimal performance. Consequently, it is critical to balance the trade-off between performance (benefits) and computational costs (budgets), giving rise to the concept of reasoning economy. In this survey, we provide a comprehensive analysis of reasoning economy in both the post-training and test-time inference stages of LLMs, encompassing i) the cause of reasoning inefficiency, ii) behavior analysis of different reasoning patterns, and iii) potential solutions to achieve reasoning economy. By offering actionable insights and highlighting open challenges, we aim to shed light on strategies for improving the reasoning economy of LLMs, thereby serving as a valuable resource for advancing research in this evolving area. We also provide a public repository to continually track developments in this fast-evolving field.

cs.CL↗

arXiv2026

Fractal transference principles for subsets of $\mathbb{N}^d$ of positive density

We establish a multidimensional fractal transference principle for digit-restricted sets associated with subsets of $\mathbb{N}^d$, extending the one-dimensional framework of Nakajima--Takahasi (Adv. Math., 2025). We develop general Hausdorff-dimension tools via the singular value potential $ϕ^s(\mathbf a)$ and the multivariate Dirichlet series $ζ_S(\boldsymbolσ) =\sum_{\mathbf a\in S}\prod_{j=1}^d a_j^{-σ_j}$. Let $s_\ast:=\inf\{0<s\le d:\sum_{\mathbf a\in S}ϕ^s(\mathbf a)<\infty\}$ and $Λ_S:=\inf\{σ_1+\cdots+σ_d:ζ_S(\boldsymbolσ)<\infty\}$. We obtain $\dim_H(\mathcal E_S)\le s_\ast$, where $\mathcal E_S\subset((0,1)\setminus\mathbb{Q})^d$ denotes the set of points whose continued-fraction digit vectors lie in $S$ and whose coordinates escape (i.e.\ $a_n(x_j)\to\infty$ for each $j$), and $s_\ast=\tfrac12Λ_S$ for uniformly $K$--balanced $S$. If $S\subset\mathbb{N}^d$ has positive upper density, the transference theorem constructs a set $E_S\subset\mathcal E_{\mathbb N^d}^{\mathrm{vec}}$ with $\dim_H E_S=d/2$; in the positive upper Banach density case we can construct $F_S\subset\mathcal E_{\mathbb N^d}^{\mathrm{vec}}$ with $\dim_H F_S=d/2$. In both cases the common digit set recovers the corresponding density of $S$. On the combinatorial side, the transference principle ensures that translation-invariant configurations forced at positive density, including multidimensional Szemerédi patterns, persist inside the induced fractal digit sets.

math.DS↗

arXiv2026

Connectedness of polynomial diagonal orbit closures for minimal nilrotations and applications

For a minimal nilrotation on a compact connected nilmanifold, we prove that the polynomial diagonal orbit closure associated with any finite family of polynomials with integer coefficients vanishing at the origin is connected. This resolves a conjecture of Glasscock, Koutsogiannis, Le, Moreira, Richter, and Robertson. Combined with their equivalence theorem, our result yields polynomial multiple recurrence in every prescribed residue class in topological dynamics, provided that the corresponding power of the transformation is minimal. Furthermore, we independently establish the measure-theoretic counterpart of this recurrence phenomenon. Finally, we construct a totally minimal nilsystem for which the lower central series identity proposed by Leibman fails.

math.DS↗

arXiv2026

Zero-Threshold Discrepancies for Multiple Correlation Sequences

We study the zero-threshold lifting problem for polynomial multiple correlation sequences with respect to the measure-theoretic pro-nilfactor. The structure theory for polynomial multiple averages implies that, at every positive threshold, positivity on the pro-nilfactor lifts to positivity in the original system, except on a set of zero upper Banach density. We demonstrate that this lifting property does not hold at the zero threshold. Specifically, we construct an ergodic system and two sets of positive measure for which the pro-nilfactor correlation is positive along a set of times with positive upper density, while the corresponding exact correlation vanishes on this set. This provides a negative answer to a question posed by Glasscock, Koutsogiannis, Le, Moreira, Richter, and Robertson. %\cite{GKLMMRR}. Additionally, we prove a corresponding rigidity property. For any ergodic system, any essentially distinct family of integer polynomials vanishing at the origin, and any tuple of non-negative bounded functions, the zero-threshold discrepancy set is not piecewise syndetic.

math.DS↗

arXiv2026

Position: Agent Should Invoke External Tools ONLY When Epistemically Necessary

As large language models evolve into tool-augmented agents, a central question remains unresolved: when is external tool use actually justified? Existing agent frameworks typically treat tools as ordinary actions and optimize for task success or reward, offering little principled distinction between epistemically necessary interaction and unnecessary delegation. This position paper argues that agents should invoke external tools only when epistemically necessary. Here, epistemic necessity means that a task cannot be completed reliably via the agent's internal reasoning over its current context, without any external interaction. We introduce the Theory of Agent (ToA), a framework that treats agents as making sequential decisions about whether remaining uncertainty should be resolved internally or delegated externally. From this perspective, common agent failure modes (e.g., overthinking and overacting) arise from miscalibrated decisions under uncertainty rather than deficiencies in reasoning or tool execution alone. We further discuss implications for training, evaluation, and agent design, highlighting that unnecessary delegation not only causes inefficiency but can impede the development of internal reasoning capability. Our position provides a normative criterion for tool use that complements existing decision-theoretic models and is essential for building agents that are not only correct, but increasingly intelligent.

cs.AI↗

arXiv2026

Learning Agent Routing From Early Experience

LLM agents achieve strong performance on complex reasoning tasks but incur high latency and compute cost. In practice, many queries fall within the capability boundary of cutting-edge LLMs and do not require full agent execution, making effective routing between LLMs and agents a key challenge. We study the problem of routing queries between lightweight LLM inference and full agent execution under realistic cold-start settings. To address this, we propose BoundaryRouter, a training-free routing framework that uses early behavioral experience and rubric-guided reasoning to decide whether to answer a query with direct LLM inference or escalate to an agent. BoundaryRouter builds a compact experience memory by executing both systems on a shared seed set and retrieves similar cases at inference time to guide routing decisions. To evaluate this method, we introduce RouteBench, a benchmark covering in-domain, paraphrased, and out-of-domain route settings. Experiments show that BoundaryRouter reduces inference time by 60.6% compared to the agent while improving performance by 28.6% over direct LLM inference, outperforming prompt-based and retrieval-only routing by an average of 37.9% and 8.2%, respectively.

cs.CL↗

arXiv2026

From Word to World: Can Large Language Models be Implicit Text-based World Models?

Agentic reinforcement learning increasingly relies on experience-driven scaling, yet real-world environments remain non-adaptive, limited in coverage, and difficult to scale. World models offer a potential way to improve learning efficiency through simulated experience, but it remains unclear whether large language models can reliably serve this role and under what conditions they meaningfully benefit agents. We study these questions in text-based environments, which provide a controlled setting to reinterpret language modeling as next-state prediction under interaction. We introduce a three-level framework for evaluating LLM-based world models: (i) fidelity and consistency, (ii) scalability and robustness, and (iii) agent utility. Across five representative environments, we find that sufficiently trained world models maintain coherent latent state, scale predictably with data and model size, and improve agent performance via action verification, synthetic trajectory generation, and warm-starting reinforcement learning. Meanwhile, these gains depend critically on behavioral coverage and environment complexity, delineating clear boundry on when world modeling effectively supports agent learning.

cs.CL↗

arXiv2026

Position: The Inevitable End of One-Architecture-Fits-All-Domains in Time Series Forecasting

Recent work has questioned the effectiveness and robustness of neural network architectures for time series forecasting tasks. We summarize these concerns and analyze groundly their inherent limitations: i.e. the irreconcilable conflict between single (or few similar) domains SOTA and generalizability over general domains for time series forecasting neural network architecture designs. Moreover, neural networks architectures for general domain time series forecasting are becoming more and more complicated and their performance has almost saturated in recent years. As a result, network architectures developed aiming at fitting general time series domains are almost not inspiring for real world practices for certain single (or few similar) domains such as Finance, Weather, Traffic, etc: each specific domain develops their own methods that rarely utilize advances in neural network architectures of time series community in recent 2-3 years. As a result, we call for the time series community to shift focus away from research on time series neural network architectures for general domains: these researches have become saturated and away from domain-specific SOTAs over time. We should either (1) focus on deep learning methods for certain specific domain(s), or (2) turn to the development of meta-learning methods for general domains.

cs.LG↗

arXiv2026

A Survey of Self-Evolving Agents: What, When, How, and Where to Evolve on the Path to Artificial Super Intelligence

Large Language Models (LLMs) have demonstrated remarkable capabilities across diverse tasks but remain fundamentally static, unable to adapt their internal parameters to novel tasks, evolving knowledge domains, or dynamic interaction contexts. As LLMs are increasingly deployed in open-ended, interactive environments, this static nature has become a critical bottleneck, necessitating agents that can adaptively reason, act, and evolve in real time. This paradigm shift -- from scaling static models to developing self-evolving agents -- has sparked growing interest in architectures and methods enabling continual learning and adaptation from data, interactions, and experiences. This survey provides the first systematic and comprehensive review of self-evolving agents, organizing the field around three foundational dimensions: what, when, and how to evolve. We examine evolutionary mechanisms across agent components (e.g., models, memory, tools, architecture), categorize adaptation methods by stages (e.g., intra-test-time, inter-test-time), and analyze the algorithmic and architectural designs that guide evolutionary adaptation (e.g., scalar rewards, textual feedback, single-agent and multi-agent systems). Additionally, we analyze evaluation metrics and benchmarks tailored for self-evolving agents, highlight applications in domains such as coding, education, and healthcare, and identify critical challenges and research directions in safety, scalability, and co-evolutionary dynamics. By providing a structured framework for understanding and designing self-evolving agents, this survey establishes a roadmap for advancing more adaptive, robust, and versatile agentic systems in both research and real-world deployments, and ultimately sheds light on the realization of Artificial Super Intelligence (ASI) where agents evolve autonomously and perform beyond human-level intelligence across tasks.

cs.AI↗

arXiv2026

On subsets of integers having dense orbits

Let $A\subset \mathbb{N}$. We say $A$ is an $R$-sequence for a given minimal system $(Y,S)$ if there is $y\in Y$ such that $\{S^ny:n\in A\}$ is dense in $Y$. Richter asked if $A$ is an $R$-sequence for all minimal equicontinuous systems implies that $A$ is an $R$-sequence for all minimal systems. In this paper, we investigate this question and related issues within the framework of totally minimal systems, including a characterization of transitive systems that are disjoint from all totally minimal systems. A dynamical system is scattering (resp. weakly scattering) if its product with any minimal (resp. minimal and equicontinuous) system is transitive. It turns out that $(X,T)$ is scattering if and only if for any transitive point $x\in X$ and any minimal system $(Y,S)$ there is $y\in Y$ such that the orbit of $(x,y)$ is dense in $X\times Y$ if and only if for each transitive point $x\in X$ and any non-empty open subset $U$ of $X$, $\{n\in \mathbb{N}:T^nx\in U\}$ is an $R$-sequence. By combining this result with earlier work of Huang and Ye, we deduce that if scattering and weak scattering are distinct properties, then both Richter's question and Katznelson's question admit negative answers.

math.DS↗

arXiv2026

CubeBench: Diagnosing Interactive, Long-Horizon Spatial Reasoning Under Partial Observations

Large Language Model (LLM) agents, while proficient in the digital realm, face a significant gap in physical-world deployment due to the challenge of forming and maintaining a robust spatial mental model. We identify three core cognitive challenges hindering this transition: spatial reasoning, long-horizon state tracking via mental simulation, and active exploration under partial observation. To isolate and evaluate these faculties, we introduce CubeBench, a novel generative benchmark centered on the Rubik's Cube. CubeBench uses a three-tiered diagnostic framework that progressively assesses agent capabilities, from foundational state tracking with full symbolic information to active exploration with only partial visual data. Our experiments on leading LLMs reveal critical limitations, including a uniform 0.00% pass rate on all long-horizon tasks, exposing a fundamental failure in long-term planning. We also propose a diagnostic framework to isolate these cognitive bottlenecks by providing external solver tools. By analyzing the failure modes, we provide key insights to guide the development of more physically-grounded intelligent agents.

cs.AI↗

arXiv2025

GenEnv: Difficulty-Aligned Co-Evolution Between LLM Agents and Environment Simulators

Training capable Large Language Model (LLM) agents is critically bottlenecked by the high cost and static nature of real-world interaction data. We address this by introducing GenEnv, a framework that establishes a difficulty-aligned co-evolutionary game between an agent and a scalable, generative environment simulator. Unlike traditional methods that evolve models on static datasets, GenEnv instantiates a dataevolving: the simulator acts as a dynamic curriculum policy, continuously generating tasks specifically tailored to the agent's ``zone of proximal development''. This process is guided by a simple but effective $α$-Curriculum Reward, which aligns task difficulty with the agent's current capabilities. We evaluate GenEnv on five benchmarks, including API-Bank, ALFWorld, BFCL, Bamboogle, and TravelPlanner. Across these tasks, GenEnv improves agent performance by up to \textbf{+40.3\%} over 7B baselines and matches or exceeds the average performance of larger models. Compared to Gemini 2.5 Pro-based offline data augmentation, GenEnv achieves better performance while using 3.3$\times$ less data. By shifting from static supervision to adaptive simulation, GenEnv provides a data-efficient pathway for scaling agent capabilities.

cs.CL↗

arXiv2025

Zoom in, Click out: Unlocking and Evaluating the Potential of Zooming for GUI Grounding

Grounding is a fundamental capability for building graphical user interface (GUI) agents. Although existing approaches rely on large-scale bounding box supervision, they still face various challenges, such as cross-platform generalization, complex layout analysis, and fine-grained element localization. In this paper, we investigate zoom as a strong yet underexplored prior for GUI grounding, and propose a training-free method, ZoomClick. By characterizing four key properties of zoom (i.e., pre-zoom, depth, shrink size, minimal crop size), we unlock its full capabilities for dynamic spatial focusing and adaptive context switching. Experiments demonstrate that our method significantly boosts the performance of both general vision-language and specialized GUI grounding models, achieving state-of-the-art results on several mainstream benchmarks; for example, UI-Venus-72B attains a 73.1% success rate on ScreenSpot-Pro. Furthermore, we present GUIZoom-Bench, a benchmark for evaluating model adaptability to zoom, aiming to inspire future research on improving zoom for further training and test-time scaling in GUI grounding tasks.

cs.CV↗

arXiv2025

Alita-G: Self-Evolving Generative Agent for Agent Generation

Large language models (LLMs) have been shown to perform better when scaffolded into agents with memory, tools, and feedback. Beyond this, self-evolving agents have emerged, but current work largely limits adaptation to prompt rewriting or failure retries. Therefore, we present ALITA-G, a self-evolution framework that transforms a general-purpose agent into a domain expert by systematically generating, abstracting, and curating Model Context Protocol (MCP) tools. In this framework, a generalist agent executes a curated suite of target-domain tasks and synthesizes candidate MCPs from successful trajectories. These are then abstracted to parameterized primitives and consolidated into an MCP Box. At inference time, ALITA-G performs retrieval-augmented MCP selection with the help of each tool's descriptions and use cases, before executing an agent equipped with the MCP Executor. Across several benchmarks GAIA, PathVQA, and Humanity's Last Exam, ALITA-G attains strong gains while reducing computation costs. On GAIA validation, it achieves 83.03% pass@1 and 89.09% pass@3, establishing a new state-of-the-art result while reducing mean tokens per example by approximately 15% relative to a strong baseline agent. ALITA-G thus provides a principled pathway from generalist capability to reusable, domain-specific competence, improving both accuracy and efficiency on complex reasoning tasks.

cs.AI↗

arXiv2025

On the Role of Preference Variance in Preference Optimization

Direct Preference Optimization (DPO) has emerged as an important approach for learning from human preferences in aligning large language models (LLMs). However, collecting human preference data is costly and inefficient, motivating methods to reduce the required annotations. In this work, we investigate the impact of \emph{preference variance} (PVar), which measures the variance in model preferences when comparing pairs of responses, on the effectiveness of DPO training. We provide a theoretical insight by establishing an upper bound on the DPO gradient norm for any given prompt, showing it is controlled by the PVar of that prompt. This implies that prompts with low PVar can only produce small gradient updates, making them less valuable for learning. We validate this finding by fine-tuning LLMs with preferences generated by a reward model, evaluating on two benchmarks (AlpacaEval 2.0 and Arena-Hard). Experimental results demonstrate that prompts with higher PVar outperform randomly selected prompts or those with lower PVar. We also show that our PVar-based selection method is robust, when using smaller reward models (1B, 3B) for selection. Notably, in a separate experiment using the original human annotations from the UltraFeedback dataset, we found that training on only the top 10\% of prompts with the highest PVar yields better evaluation performance than training on the full dataset, highlighting the importance of preference variance in identifying informative examples for efficient LLM alignment.

cs.CL↗

arXiv2025

ReasonFlux-PRM: Trajectory-Aware PRMs for Long Chain-of-Thought Reasoning in LLMs

Process Reward Models (PRMs) have recently emerged as a powerful framework for supervising intermediate reasoning steps in large language models (LLMs). Previous PRMs are primarily trained on model final output responses and struggle to evaluate intermediate thinking trajectories robustly, especially in the emerging setting of trajectory-response outputs generated by frontier reasoning models like Deepseek-R1. In this work, we introduce ReasonFlux-PRM, a novel trajectory-aware PRM explicitly designed to evaluate the trajectory-response type of reasoning traces. ReasonFlux-PRM incorporates both step-level and trajectory-level supervision, enabling fine-grained reward assignment aligned with structured chain-of-thought data. We adapt ReasonFlux-PRM to support reward supervision under both offline and online settings, including (i) selecting high-quality model distillation data for downstream supervised fine-tuning of smaller models, (ii) providing dense process-level rewards for policy optimization during reinforcement learning, and (iii) enabling reward-guided Best-of-N test-time scaling. Empirical results on challenging downstream benchmarks such as AIME, MATH500, and GPQA-Diamond demonstrate that ReasonFlux-PRM-7B selects higher quality data than strong PRMs (e.g., Qwen2.5-Math-PRM-72B) and human-curated baselines. Furthermore, our derived ReasonFlux-PRM-7B yields consistent performance improvements, achieving average gains of 12.1% in supervised fine-tuning, 4.5% in reinforcement learning, and 6.3% in test-time scaling. We also release our efficient ReasonFlux-PRM-1.5B for resource-constrained applications and edge deployment. Project: https://github.com/Gen-Verse/ReasonFlux

cs.CL↗

arXiv2025

TreeBoN: Enhancing Inference-Time Alignment with Speculative Tree-Search and Best-of-N Sampling

Inference-time alignment enhances the performance of large language models without requiring additional training or fine-tuning but presents challenges due to balancing computational efficiency with high-quality output. Best-of-N (BoN) sampling, as a simple yet powerful approach, generates multiple responses and selects the best one, achieving improved performance but with a high computational cost. We propose TreeBoN, a novel framework that integrates a speculative tree-search strategy into Best-of-N (BoN) Sampling. TreeBoN maintains a set of parent nodes, iteratively branching and pruning low-quality responses, thereby reducing computational overhead while maintaining high output quality. Our approach also leverages token-level rewards from Direct Preference Optimization (DPO) to guide tree expansion and prune low-quality paths. We evaluate TreeBoN using AlpacaFarm, HH-RLHF, UltraFeedback, GSM8K, and TutorEval datasets, demonstrating consistent improvements. Specifically, TreeBoN achieves the highest win rate of 65% on TutorEval and around 60% win rates across other different datasets, outperforming standard BoN with the same computational cost and showcasing its scalability and alignment efficacy.

cs.CL↗

arXiv2025

Physics Supernova: AI Agent Matches Elite Gold Medalists at IPhO 2025

Physics provides fundamental laws that describe and predict the natural world. AI systems aspiring toward more general, real-world intelligence must therefore demonstrate strong physics problem-solving abilities: to formulate and apply physical laws for explaining and predicting physical processes. The International Physics Olympiad (IPhO)--the world's most prestigious physics competition--offers a rigorous benchmark for this purpose. We introduce Physics Supernova, an AI agent system with superior physics problem-solving abilities that match elite IPhO gold medalists. In IPhO 2025 theory problems, Physics Supernova attains 23.5/30 points, ranking 14th of 406 contestants and surpassing the median performance of human gold medalists. We extensively analyzed Physics Supernova's capabilities and flexibility across diverse physics tasks. These results show that principled tool integration within agent systems can deliver competitive improvements in solving challenging science problems. The codes are available at https://github.com/CharlesQ9/Physics-Supernova.

cs.AI↗