arXiv subjects

Yifei Wang

Publications and source records attributed to Yifei Wang.

At least 19 recordsLinked to original sources

arXiv2026

Route Before Response: Fail-Closed Routing for LLM-Enabled Customer-Service Software

LLM-enabled customer-service software in financial institutions must choose a downstream service before responding. Classifiers and LLM routers can identify a relevant label or destination, but a composed application must still verify dispatch permission and entity registration. We study this integration problem as boundary-aware pre-response routing. We present RouteBR, a controller in which one model call proposes a primary service action and registered object, then deterministic code checks identifiers and versioned rules before dispatch or a specified non-dispatch outcome. Lower-ranked candidates cannot replace a blocked primary route. We construct 600 privacy-preserving scenarios from recurring confidential-request patterns at a real financial institution; three engineers jointly reviewed the scenarios and labels. On the 200-case CONFIRM MAIN split, RouteBR reaches 97.5% final-path exact match. A matched one-call baseline given the same rules reaches 95.5%, whereas a diagnostic baseline without the rule program obtains 94.0% action accuracy but only 41.0% final-path exact match. Across 166 non-dispatch MAIN and FLIP cases, RouteBR produces no false dispatch. The artifact pairs all 1,960 prediction rows with raw receipts while excluding gold and reviewer fields from model requests. These results support separating model proposals from code-owned, replayable execution. Our implementation is available at https://github.com/YgcCoder/RouteBR.

cs.SE↗

arXiv2026

PharmAgents2: Towards Autonomous End-to-End Drug Discovery

Drug discovery requires continuous scientific decision-making across a vast research space. We introduce PharmAgents2, an autonomous system for end-to-end drug discovery that combines knowledge-enhanced decision-making with organized large-scale exploration. Paper2Skill extracts transferable scientific decision knowledge from public sources and, together with expert-developed skills, supports research decisions. Self-Scaling Discovery expands exploration through parallel agents inheriting research context and decision-making capabilities across targets, scientific hypotheses, and discovery strategies. We introduce PharmAgents-Bench v0.1, comprising 121 questions across nine decision scenarios and three end-to-end tasks assessed with expert-defined process rubrics. PharmAgents2 outperforms evaluated baselines in most decision categories and achieves the highest process scores on all three end-to-end tasks. In three real-world studies, autonomous investigations generated findings subsequently supported by human-executed experiments. For LpxH, PharmAgents2 hypothesized a previously unreported allosteric site and screened compounds against it, identifying a novel inhibitor scaffold with experimentally confirmed inhibitory activity and binding. For JAK1, it prioritized a covalent inhibition strategy and nominated compounds with experimentally demonstrated activity. Starting from a known bioactive molecule, Self-Scaling Discovery explored candidate targets across the proteome and identified a new binding target, with the interaction subsequently confirmed experimentally. PharmAgents2 represents a prototype of a virtual pharmaceutical company, marking a step toward autonomous end-to-end drug discovery.

cs.AI↗

arXiv2026

Magna: simulating magic state factories as cheaply as Clifford circuits

Classical simulation of magic state factories, including distillation, cultivation, and code switching, is challenging because of their substantial non-Clifford content. We introduce $\texttt{magna}$ (magic-state-factory assessment via global network algebra), a classical simulator that simulates such protocols as cheaply as Clifford circuits. $\texttt{magna}$ expresses the figures of merit of a protocol, namely the acceptance rate and the fidelity of the accepted states, as tensor networks, which it reduces globally by exploiting a structural correspondence with transversal gates on quantum codes. $\texttt{magna}$ further supports feedforward conditioned on arbitrary functions of mid-circuit measurement outcomes, including decoder outputs. We simulate Bravyi--Haah distillation with up to 100 output magic states at a cost that does not grow exponentially in the number of outputs, and achieve state-of-the-art sampling times for distance-5 magic state cultivation on both color codes and RP$^2$ surface codes. Finally, we validate nonlinear feedforward on a logical-level distillation protocol with 15 rounds of decoder-dependent feedforward. $\texttt{magna}$ aims to contribute to the key infrastructure for the large-scale evaluation and optimization of magic state factories.

quant-ph↗

arXiv2026

Toric codes all the way down: decoupling translation-invariant topological CSS codes in higher dimensions

Topological Calderbank-Shor-Steane (CSS) codes in more than two spatial dimensions offer features useful for fault-tolerant quantum computation, such as transversal non-Clifford gates and single-shot decoding, but a systematic understanding of them is still lacking. One approach is to decouple a code: map it to a stack of higher-form toric codes by a finite-depth local CSS-preserving Clifford circuit, possibly after adding ancillas. For translation-invariant $p$-ary CSS codes, where $p$ is prime ($p=2$ for qubits), we require the circuit to commute with the translations of a superlattice. In two dimensions, such a decoupling exists once these translations act trivially on every anyon type. In higher dimensions, excitations are sorted by their dimension into superselection layers, which can remain glued together even when translations act trivially on each layer. We show that this gluing is the only remaining obstruction: a code can be decoupled on a superlattice exactly when, after coarse-graining to it, translations act trivially on every layer and the gluing vanishes. Moreover, for a code with finitely many excitation types in each layer and $m$ nonzero layers, once translations act trivially on every layer, further coarse-graining by a factor of $p^{m-1}$ in every direction removes the aforementioned gluing. Physically, the same criterion determines when the code encodes the maximal number of logical qudits on a finite torus. Finally, we give an explicit algorithm that constructs the decoupling as a circuit of SUM (qudit CNOT) and SWAP gates.

quant-ph↗

arXiv2026

Bug2Scenario: Root-Cause-Preserving Test Scenario Amplification for Autonomous Driving Software

Simulation-based fuzzing exposes autonomous driving system (ADS) failures by repeatedly mutating and executing scenarios, but converting one failure into reusable regression scenarios remains difficult. Such variants test whether the mechanism persists after software changes, for example by varying weather or actor placement while retaining a collision's actor-route conflict. We present Bug2Scenario, a root-cause-guided post-fuzzing framework. From scenario data, traces, and oracle evidence, it extracts a root-cause pattern. An LLM proposes semantic variations over actors, missions, routes, and weather. Deterministic rules construct DriveFuzz-compatible scenarios. Execution evidence and manual review assess preservation. Separating these responsibilities prevents fluency, syntax, or recurring labels from masquerading as causal evidence. Under matched seeds, six LLM backbones, candidate budget, and execution pipeline, Bug2Scenario improves strict preservation by 6.7 percentage points (p=0.104; not significant) and boundary-inclusive preservation by 10.6 points (p=0.011). Feedback helps some initially unsuccessful seeds yield a strictly preserved candidate. Seed-level stopping then avoids further API calls and simulator runs. Remaining failures reflect actor-placement constraints, simulator instability, and incomplete or conflicting causal evidence. Bug2Scenario thus produces reliable ADS regression assets. Our implementation is available at https://github.com/YgcCoder/Bug2Sce.

cs.SE↗

arXiv2026

Design of an Underwater Vehicle With Orientation-Adjustable Thrusters and Adaptive Tracking Control

Unmanned underwater vehicles (UUVs) are essential for marine exploration and research. However, conventional designs often struggle with limited maneuverability in complex and dynamic underwater environments. This paper presents an orientation-adjustable thruster UUV (OAT-UUV) with vector thrusters, enabling full six-degree-of-freedom (6-DOF) motion. and composite maneuvers. To overcome challenges associated with uncertain model parameters and environmental disturbances, a vision-based tracking control system based on a novel feedforward adaptive model predictive controller (FF-AMPC) is proposed to achieve robust object tracking, which integrates realtime state feedback with adaptive parameter updating. Extensive experiments, including closed-loop tracking and composite motion tests in real-world environment, have been conducted. The obtained results demonstrate that the orientation-adjustable propulsion design significantly improves the underwater maneuverability, while the FF-AMPC controller achieves a 49.1% hovering improvement compared to conventional controller.

cs.RO↗

arXiv2026

The Cryptocurrency Trend Factor Revisited: Predictive Information, Portfolio Construction, and Pricing Claims

We reassess five economic claims about the cryptocurrency trend factor (CTREND): cross-sectional return predictability, high risk-adjusted performance, incremental information beyond market, size, and momentum, the ability to replace momentum in factor models, and robustness among large and liquid coins and after trading costs. Our reconstruction and code audit identify issues in information timing, technical indicators, forecast coverage, portfolio formation, and return accounting. We first correct future-information use and related settlement rules, then implement broader corrections to indicators, forecast availability, and portfolio accounting. Without additional coin exclusions, the first stage changes CTREND's mean weekly gross return from 4.07% to 4.25%, annualized Sharpe ratio from 2.08 to 1.48, and weekly alpha relative to published market, size, and momentum factors from 2.77% to 3.06%. After comprehensive correction, these become 0.05%, 0.01, and -1.38%; the alpha is no longer significant. Replacing momentum with CTREND yields an advantage in maximum squared Sharpe ratios of 0.058 in the original implementation, 0.024 after the first-stage correction, and -0.0185 after comprehensive correction, with a statistically significant overall decline. Mean weekly gross returns for the largest 100 coins change from 2.93% to 3.99% to 1.00%, while mean weekly net returns for the full long-short portfolio under the lowest assumed trading costs change from 2.93% to 3.35% to -0.87%. These findings show that reported performance and factor-substitution gains depend materially on implementation and sample composition. We distinguish predictive information, an independent return premium, and the incremental value of replacing an existing factor as separate economic claims.

cs.CE↗

arXiv2026

The nonexistence of nine-point crescent configurations

A planar crescent configuration consists of $n$ points, no three collinear and no four concyclic, whose distinct distances have multiplicities $1,2,\ldots,n-1$. We prove by exact computation that no such configuration exists for $n=9$. The three pairs carrying the two least frequent distances yield a six-point subset with at most six distances. Rational linear algebra on quartic polynomials separates the possible six-point distance patterns into those with certified finite lists of normalized distance tuples and a residual family. Compatibility of overlapping subsets forces at least one core into the first family, without classifying the residual realizations. The remaining three points are then obtained by rational constructions over the field of the squared core distances. The multiplication matrices used to enumerate the core distances have order at most twelve. Exhaustive label assignments and exact algebraic comparisons exclude all remaining extensions.

math.MG↗

arXiv2026

Classification of crescent configurations

Let $n$ points be in crescent configuration in $\mathbb{R}^d$ if they lie in general position in $\mathbb{R}^d$ and determine $n-1$ distinct distances, such that for every $1 \leq i \leq n-1$ there is a distance that occurs exactly $i$ times. For the planar problem, Erdős conjectured that there exists $N$ such that no crescent configuration exists on $n$ points whenever $n\geq N$. In the same paper, he also gave constructions for $n$ up to 8, with $n=5$ due to Pomerance and $n=7,8$ due to Palásti, but nothing is yet known for $n \geq 9$. Most recently, Burt et al. proved that a crescent configuration on $n$ points exists in $\mathbb{R}^{n-2}$ for $n \geq 3$. In this paper, we classify planar crescent configurations by their complete distance-equality patterns, up to relabeling the vertices. We restrict attention to the strict distance classes defined below. There are three such classes on four vertices, all of which admit planar crescent realizations. On five vertices, there are $49$ strict classes, exactly $31$ of which admit planar crescent realizations. On six vertices, there are $6837$ strict classes, exactly six of which admit planar crescent realizations. The classification combines finite enumeration with exact algebraic verification of realizations and nonexistence. The distance-geometric criteria also apply in higher dimensions.

math.CO↗

arXiv2026

Strong yet Controllable LLM Abstention without Learning to Abstain

Large language models (LLMs) can reduce hallucination by abstaining from questions they would likely get wrong. Abstention can be implemented outside the model (selective answering) using a confidence threshold that controls the answer rate (coverage), or learned inside it (abstention training), which can perform better but fixes coverage once trained. To compare the two, we match their coverage and decompose risk (error rate) into answering ability and distinguishing ability, i.e., how well the model answers and how accurately it abstains, respectively. The decomposition shows that abstention training performs better mainly by answering better: with strong confidence estimators, the two have nearly equal distinguishing ability. This motivates Train for Accuracy, then Post-hoc Select (TAPS), a modular approach that keeps coverage controllable, needs only standard post-training and an off-the-shelf confidence estimator, and benefits from advances in either component. Across three LLMs and three factual QA benchmarks, TAPS with a probing-based confidence estimator attains 1.5--9.4 percentage points lower mean risk than three representative abstention training methods at matched coverage. Within TAPS, accuracy training often weakens the model's own confidence signals, yet refitting the estimator largely preserves distinguishing ability. Together, these results establish TAPS as a strong yet controllable baseline and raise the bar for future abstention training: sacrificing control over coverage should yield gains beyond what can be achieved by optimizing answering and abstaining separately.

cs.CL↗

arXiv2026

Semiparametric inference for a Target Outcome Leveraging Heterogeneous External Summaries of Auxiliary Outcomes

The integration of external information can substantially improve the efficiency of statistical inference in an internal study, but is often complicated by between-study heterogeneity. Additional challenges arise when external studies examine auxiliary outcomes rather than on the target outcomes and provide only summary statistics. To address these challenges, we propose a semiparametric framework that incorporates external summary statistics into the internal analysis via the generalized method of moments. To accommodate between-study heterogeneity, the external summaries are formulated within a random-effects model centered on the internal parameter estimates and subsequently embedded into the internal estimating equations. We establish the asymptotic properties of the proposed estimator. While uncertainty from the first-stage estimation vanishes as the number of external studies increases, it remains non-negligible where the number of external studies is moderate. In this setting, we show that a hybrid bootstrap procedure correctly propagates this uncertainty and consistently estimates the asymptotic covariance of the proposed estimator, thereby ensuring valid inference. We further derive theoretical efficiency bounds and demonstrate robust finite-sample performance through simulation studies. In an application to colorectal cancer tumor sequencing data, we show that integrating external information of auxiliary outcomes reveals molecular subtype-specific etiologic associations that would not be detected using the internal data alone.

stat.ME↗

arXiv2026

QUEST: Training Frontier Deep Research Agents with Fully Synthetic Tasks

Deep research agents extend the role of search engines from retrieving keyword-matched pages to synthesizing knowledge, fundamentally changing how humans interact with information. However, frontier systems remain proprietary, while existing open agents often generalize poorly across different task types, leaving unclear how to train a broadly capable deep research agent. We release QUEST, a family of open models (ranging from 2B to 35B) that serve as general-purpose deep research agents designed to handle a wide range of long-horizon search tasks, with strong capabilities in fact seeking, citation grounding, and report synthesis. To build QUEST, we propose an effective training recipe combining mid-training, supervised fine-tuning, and reinforcement learning. Central to this recipe is a curated data synthesis pipeline based on unified rubric trees, which applies to different task types and enables synthesizing training data with verifiable rewards without human annotation. In addition, QUEST incorporates a built-in context management mechanism that enables effective long-horizon reasoning and knowledge synthesis. Using only 8K synthesized tasks, QUEST approaches or even surpasses frontier closed-source agents across eight deep research benchmarks spanning diverse task types, and achieves the best overall performance among recent open-weight agents. We released everything: models, data, and training scripts.

cs.CL↗

arXiv2026

Solvent: Long-Horizon Reasoning for Multi-Step Retrosynthesis

Multi-step retrosynthesis requires recursively transforming a target molecule through successive reaction steps into purchasable building blocks, and recent work has begun to draw on large language model (LLM) reasoning to tackle the task. In its latest evolution, LLM reasoning has entered the long-horizon regime: models pursue objectives across interdependent steps, interact with environments, and adapt to feedback. This shift has markedly advanced LLM capabilities, as illustrated by breakthroughs in AI-driven mathematical research. At this frontier, we introduce Solvent, an 8B LLM that successfully extends long-horizon reasoning to multi-step retrosynthesis. Through a two-round post-training curriculum, Solvent learns to interleave iterative synthesis planning, external availability verification, and adaptive route revision. Experiments show that, following a $2.1\times$ performance boost from RL, Solvent scores 61.86% across four benchmarks, surpassing mid-scale language models and prior search-based retrosynthesis systems, and rivaling frontier trillion-parameter LLMs.

cs.AI↗

arXiv2026

When Is a Memory Safe Enough to Act On? Action-Conditioned Verification for Long-Lived Agents

Long-lived agents act on memory carried across sessions, and many of their actions are irreversible, so a stale memory can turn directly into harm. Existing work improves the memory itself --- retrieving, refreshing, or reconciling entries --- yet whether an entry must be re-checked before acting is not a property of the entry: the same remembered address can be used unchecked for a routine shipment but should be re-confirmed before a valuable one. We therefore study \emph{action-conditioned memory verification}: given a candidate action, decide which of its memory premises to re-verify and when to proceed or abstain, balancing execution risk, abstention, and verification cost. We formulate this as a sequential stopping problem, in which a verification is worth its cost only if it can change what the agent does next; under suitable verifier reliability and cost conditions, increasing action consequence can move the same memory through proceed, verify, and direct-abstention regimes. Building on this formulation, \textsc{AMV-Guard} is a practical verification gate that extracts the premises a candidate action rests on, calibrates beliefs about whether they are wrong, and selects checks by bounded value of information (VOI). On ToolSandbox, relative to a static policy given the same inputs, it lowers mean verification cost by 33.9\% while keeping 100\% safe-proceed recall and 0\% Gate harmful, and raises Proceed coverage from 38.83\% to 44.17\%; in a gate-core transfer to AppWorld, which carries over the stopping/VOI core but not the full structured pipeline, it lowers verification cost by 21.3\% at a small cost in safe-proceed recall and task success. These results suggest that verification effort should be set by what an agent is about to do, not only by how reliable a memory appears.

cs.MA↗

arXiv2026

MCSplat: Multi-View Photometric and Geometric Consistent Feed-Forward Gaussian Splatting for Driving Scenes

Driving-scene reconstruction via feed-forward Gaussian splatting from sparse multi-view images remains challenging due to two factors. First, cross-camera photometric inconsistency, caused by variations in exposure, ISP, and illumination, can be incorrectly absorbed into Gaussian geometry. Second, unreliable local geometric supervision arises from normal-based constraints near depth discontinuities. To address these issues, we present \textbf{MCSplat}, a multi-view consistent feed-forward Gaussian splatting framework for driving scenes. MCSplat constructs a canonical appearance space from depth-guided multi-camera correspondences to decouple geometry learning from camera-specific appearance. A multi-scale Bilateral Grid Transformer restores raw camera appearance from canonical renderings, enabling photometric supervision without contaminating Gaussian geometry. To improve local geometric reliability, we estimate Fisher-information-based uncertainty from photometric sensitivity and use it to adaptively weight surface-oriented self-supervision. Extensive experiments demonstrate improved reconstruction quality and generalization across driving datasets. Geometry-oriented evaluations further reveal enhanced cross-view depth consistency and local surface-normal agreement, while qualitative results show fewer geometric artifacts and more coherent structures across overlapping views.

cs.CV↗

arXiv2026

Seeing Through Sparse Signals: CS-Enhanced Sensing With QoS Guarantees for 6G ISAC

In integrated sensing and communication (ISAC) systems, communication and sensing share time-frequency resources, creating a fundamental trade-off between communication quality of service (QoS) and quality of sensing (QoSe). Reusing 5G new radio (NR) reference signals (RSs) enables standard-compliant sensing, but their inherent sparsity leads to undersampled channels and strong sidelobes that degrade weak-target detection. Compressed sensing (CS) can mitigate these effects, yet existing methods either overlook communication QoS constraints or rely on reconstruction algorithms that are step-size-sensitive or employ global thresholding, leading to weak-target masking in multi-target scenarios. The multi-target detection problem, therefore, requires a joint design of resource scheduling and CS-based reconstruction. In this work, we investigate the impact of two scheduling schemes, namely QoS-guaranteed scheduling and proportional-fairness (PF)-based scheduling, on the sensing-communication trade-off. To further enhance sensing performance, we employ a 2D max-median complex approximate message passing (2D-mmCAMP) algorithm that performs adaptive row- and column-wise thresholding to suppress sidelobes while preserving weak-target responses in multi-target scenarios. Simulations demonstrate an improvement of >= 16% in weak-target detection over baseline reconstruction algorithms while meeting downlink (DL) QoS constraints, validating the effectiveness of the proposed NR-compliant ISAC scheme.

eess.SP↗

arXiv2026

MemoryVLN: From Ego-Aware Memory Representation to Cross-Context Memory Reuse in Vision-Language Navigation

Large vision-language models (LVLMs) have shown strong potential for embodied vision-language navigation (VLN), yet their performance degrades on long-horizon trajectories due to context-length limits, redundant historical observations, and unstable spatial alignment across time. These issues make it difficult to both compress long navigation histories and reliably reuse past spatial experience. To address these limitations, we propose \textbf{MemoryVLN}, a unified memory-driven navigation system built around memory construction and management. During memory construction, \emph{Long Sequence Compression} employs forgetting-curve-inspired temporal sampling to preserve dense recent evidence and sparse yet representative distant observations under a limited token budget. \emph{Ego-Aware Memory Representation} further integrates 2D semantic tokens, 3D geometric tokens, and 6D agent poses, allowing each memory to encode both what was observed and where it was observed. During memory management, \emph{Hierarchical Memory Bank} maintains in-episode and context-level memories through cross-context warm-start and confidence-controlled write-back. \emph{Memory Enhancement} then retrieves the Top-$K$ relevant memories according to the current navigation state and fuses them with the current visual tokens. Together, these mechanisms transform raw navigation history into compact, spatially grounded, selectively retrievable, and reusable navigation experience. Extensive experiments on R2R-CE and RxR-CE demonstrate consistent improvements over the state-of-the-art baselines by a substantial margin (i.e., 2\% and 6.5\% in successful rate).

cs.RO↗

arXiv2026

Interactive Navigation Beyond Reachability: Benchmark and Agentic Framework

Existing embodied navigation tasks commonly assume that traversability and object states remain fixed during an episode. Yet reaching a goal may require opening a door to access a room and then opening a container to reveal a hidden target. We formulate interactive navigation beyond reachability, unifying channel-induced changes in spatial connectivity, container-induced changes in target visibility, and their sequential dependencies. We introduce MultINav-Bench, a 3K-episode ObjectGoal benchmark in MolmoSpaces spanning channel, container, and mixed tasks, with interaction-necessity annotations covering required and unnecessary episodes. Its data generation pipeline automates task construction across multiple task types. The evaluation protocol separately measures navigation success, interaction effectiveness, and cost. We further propose MultINav-Agent, a three-stage agentic framework built on a persistent, state-conditioned interaction graph. Semantic interactive perception grounds entities and interaction states; a vision-language model selects task-relevant actions based on expected changes in connectivity and target observability; and post-interaction verification updates the graph to guide subsequent navigation, exploration, or interaction. Experiments on MultINav-Bench show that MultINav-Agent consistently outperforms adapted baselines across channel, container, and mixed tasks.

cs.RO↗