arXiv:2606.17057v1 Announce Type: new Abstract: Although Knowledge Editing provides an efficient mechanism for updating the knowledge of Multimodal Large Language Models (MLLMs), we find that current paradigms still suffer from an important yet remain underexplored issue : editing decoupling failure, where entity-related knowledge can be updated when the model is triggered by multimodal inputs (text--image query pairs), however, it often reverts to outdated pre-edit facts when the paired inputs are split into unimodal ones. Our in-depth empirical analysis reveals that the entity knowledge in MLLMs is not stored as a unified representation, but is instead distributed across disentangled modality-specific pathways. As a result, updates biased toward multimodal queries fail to propagate effectively to unimodal circuits. To bridge this gap, we propose DECODE, which explicitly disentangles and localizes modality-specific neuron groups for targeted knowledge. Extensive experiments demonstrate that DECODE consistently achieves effective knowledge updates under different modality triggers, thereby mitigating editing decoupling failures.
Science Journals
arXiv:2606.17060v1 Announce Type: new Abstract: Waveform cross-correlation (WCC), when used for detection at seismic arrays, allows for a reduction in threshold by an order of magnitude compared to beamforming. An additional benefit of WCC detection is the significant suppression of signals unrelated to the studied area. This effect simplifies and improves the accuracy of local phase association with event hypotheses located near corresponding master events. All the advantages of WCC-based processing are best utilized in the detailed study of low-magnitude seismicity prior to mega-earthquakes. For the Sea of Okhotsk Mw8.3 earthquake that occurred on May 24, 2013, WCC processing detected the onset and subsequent evolution of intensive, low-magnitude seismicity during the five days leading up to the mainshock. The Kamchatka earthquake on July 29, 2025, is analyzed using a similar WCC-based processing extended by a detailed study of the recurrence curves of the cross-correlation bulletin (XSEL). The transition to the coseismic phase was marked by a progressive increase in the intensity of seismic activity with growing magnitude. These XSEL events fell below the detection level of the International Data Centre, which utilizes the same data stream from the International Monitoring System.
arXiv:2606.17085v1 Announce Type: new Abstract: The Wigner representation for quantum mechanics of particles is generalized to Bose fields. The standard Hilbert space quantization becomes, via the Weyl transform, a quantization method that consists of adding a Gaussian zeropoint field distribution to the vacuum. I comment on the possible advantages of the method in order to study quantum fields in curved spaces. I study a unified formulation of non-relativistic quantum electrodynamics in the Weyl-Wigner formalism, in terms of (classical-like) c-numbers.
arXiv:2606.17064v1 Announce Type: new Abstract: High performance computers produce extreme-scale data sets that require sampling or compression if they are to be used to their full potential. Existing data compression techniques typically exploit features such as sparsity in the data, homogeneity in the data, or {\it a priori} knowledge of what subsets of data are of most interest. Fluid dynamics data in general do not exhibit these features and so are attractive test beds for generic compression techniques that are objective, robust, and tuneable with respect to information lost due to compression. Presented here is a method based on tensor networks, specifically matrix product states or tensor trains, that meets these requirements. The method is demonstrated for compression in one-dimension and is extensible to higher dimensionality. Lossless compression is demonstrated for random Fourier series for sufficiently high bond dimension of the tensor network, with the memory required to store the tensor network scaling directly proportional to the bond dimension. The lossy compression exhibited at lower bond dimension can be well within the relative error of many fluid simulations. The compression algorithm is tested for the time evolution of Burger's equation with excellent results. We additionally demonstrate the capability to perform computations in the compressed form through a tensor network periodic convolution that can be orders of magnitude faster than using fast Fourier transforms and the convolution theorem. In addition to being an attractive method for working with data sets generated by existing computers, the tensor network methods utilised are directly translatable to the emerging paradigm of quantum computing.
arXiv:2606.17066v1 Announce Type: new Abstract: In this work, we show that the Fredholm integral equations underlying the distribution of relaxation times (DRT), the distribution of capacitive times (DCT), and related frameworks share a common mathematical structure, namely that of a Mellin convolution. This comes from the fact that all standard immittance (impedance or admittance) kernels depend on the product $\omega\tau$ rather than on $\omega$ and $\tau$ independently. Exploiting this structure, we derive an exact algebraic inversion formula in Mellin space that converts the deconvolution problem into a closed-form relation between the Mellin transform of the measured immittance and that of the unknown distribution function. The framework is validated analytically on a set of examples including the constant phase element (CPE), the Davidson-Cole (DC) model, and the finite-length Warburg model with blocking boundary conditions. It is also validated numerically using the fast Mellin transform via the fast Fourier transform algorithm for both the CPE and the DC model, including their DRT and DCT recovery under clean and noisy conditions. The approach unifies the impedance- and admittance-based inversions under a single spectral framework, and provides a new approach for the characterization of electrochemical systems from immittance data.
arXiv:2606.17634v1 Announce Type: new Abstract: Evaluating large language models (LLMs) is important for understanding their capabilities, comparing competing systems, and supporting the deployment of reliable models in practice. For open-ended tasks, pairwise evaluation has become a popular paradigm, in which two responses to the same prompt are compared and the resulting judgments are aggregated into an overall ranking. A central challenge of this paradigm is intransitivity: the induced comparison outcomes may fail to support any coherent global ranking. For example, one may observe cyclic preferences such as $A \succ B \succ C \succ A$, or inconsistencies involving ties such as $A \equiv B\equiv C\neq A$. Such contradictions make the resulting leaderboard unstable and challenging to interpret. In this paper, we propose a prompt perturbation framework for improving the consistency of pairwise LLM evaluation. Our approach generates perturbed variants of each prompt, uses the resulting comparison graphs to identify and filter out structurally inconsistent comparison patterns, and then applies standard ranking methods to the filtered comparisons. A key feature of the proposed framework is that graph-level structural consistency is incorporated explicitly into the evaluation pipeline before ranking aggregation. This provides a simple and principled way to reduce cyclic inconsistencies and improve the reliability of LLM rankings.
arXiv:2606.17642v1 Announce Type: new Abstract: Financial multimodal reasoning requires agents to coordinate numerical computation, retrieval, visual interpretation, and temporal grounding across heterogeneous evidence sources. Existing tool-augmented agents improve execution fidelity, yet remain largely stateless across episodes, repeatedly rediscovering reasoning strategies and failure patterns. In high-stakes financial settings, this leads to unreliable tool routing, noisy retrieval, and hallucination-prone reasoning. We present FinAcumen, a financial reasoning agent framework centered on selective experience memory for tool-augmented multimodal reasoning. FinAcumen accumulates financially grounded reasoning experience from prior trajectories, distilling successful strategies and failure-derived cautionary rules into a persistent memory bank. During inference, retrieved experiences condition reasoning only when semantic relevance exceeds a calibrated threshold, while irrelevant memory is explicitly suppressed through a fallback mechanism. A deterministic financial tool environment further grounds numerical computation, retrieval, visual decoding, and answer verification.Across four financial multimodal reasoning benchmarks, FinAcumen consistently improves a frozen 8B vision-language model over finance-specialized models and approaches leading proprietary general-purpose models. Further analysis shows that selective experience activation improves reasoning reliability under retrieval uncertainty. Our code is anonymously available at https://anonymous.4open.science/r/FinAcumen
arXiv:2606.17972v1 Announce Type: new Abstract: Self-supervised DINO models provide strong transferable visual representations, yet applying them directly to image segmentation remains challenging. Existing approaches commonly rely on heavy decoders with complex upsampling, introducing substantial parameter and computational overhead. We observe that introducing scale into DINO features is far more critical than increasing decoder capacity. In this work, we present SegDINO, an efficient segmentation framework that integrates a DINOv3 backbone with lightweight scale modeling. SegDINO introduces Token Pyramid Adaptation (TPA) to reorganize intermediate DINO features into a pseudo multi-scale hierarchy, and Scale-Aware Decoding (SAD) for efficient intra-scale refinement and top-down multi-scale propagation. We further curate PanCT, a new CT dataset containing 284 patients with expert-annotated pancreatic tumors, to assess SegDINO's ability to handle difficult small-lesion cases. Extensive experiments on PanCT and three public benchmarks demonstrate that SegDINO achieves state-of-the-art results with high efficiency. The code is available at https://github.com/script-Yang/segdino_v2.
arXiv:2606.17978v1 Announce Type: new Abstract: Trajectory similarity is a fundamental task in analyzing mobility patterns, essential for applications such as route pattern extraction, mobility prediction, and anomaly detection. Traditional distance-based measures for computing similarity incur high computational cost, driving the adoption of lightweight learning-based approaches. Supervised methods rely on extensive labels derived from traditional distance measures and often reproduce these metrics, which limits generalization. While self-supervised learning addresses this issue through contrastive learning, it lacks a unified framework, making it difficult to compare deep learning (DL) models for consistent trajectory representation. Accordingly, this paper presents MoCo-AIS, a unified framework for learning vessel trajectory embeddings based on the Momentum Contrast (MoCo) paradigm, which formulates similarity learning through positive and negative trajectory pairs. Within this framework, we evaluate a diverse set of leading DL models on large-scale, real-world vessel-tracking AIS datasets that capture diverse navigation behaviors and operating conditions. Results demonstrate that our framework significantly improves similarity learning over existing baselines, while providing a benchmarking platform for evaluating trajectory representation models.
arXiv:2606.03177v2 Announce Type: replace Abstract: Human demonstrations provide strong priors for robot manipulation, yet it is non-trivial to transfer them to execute on real robots due to the kinematic gap. In dexterous manipulation, it remains challenging to track long-horizon, contact-rich sequences even in simulators: a reference-tracking policy must keep objects on their target trajectories while preserving demonstrated joint motion and contact timing. Existing approaches often rely on hand-crafted reward tuning that require per-sequence tuning and break under limited interaction budgets. We introduce ConTrack, a reinforcement learning (RL) framework that scales with tracking data. ConTrack treats object tracking as a constraint and allocates remaining control authority to motion fidelity, which allows it to adapt task--style trade-offs online using a dual-variable update. In addition, ConTrack also stabilizes long-horizon learning with an adaptive mid-trajectory reset library that reuses policy-reachable simulator states. Our qualitative and quantitative results in simulation tracking and real robot demonstrate that ConTrack improves success and object pose accuracy significantly over prior arts while preserving joint and contact fidelity. Website: https://www.lyt0112.com/projects/ConTrack.
arXiv:2606.18190v1 Announce Type: new Abstract: Multi-stage cyberattacks span system, network, and browser logs. Detecting them requires correlating events across all three sources. Machine learning methods can learn these cross-source patterns, but they need labeled multi-source data. Existing public datasets fall short. Network-only datasets such as CICIDS and UNSW-NB15 miss host and browser activity. Host-focused datasets such as LMDG and CICAPT-IIoT lack browser telemetry. ATLAS includes all three sources but labels events only as malicious or benign, without MITRE Adversarial Tactics, Techniques, and Common Knowledge (ATT&CK) technique granularity. No public dataset combines all three sources with per-entry ATT&CK technique labels. We close the gap by building a multi-source log dataset of 870 sessions (70 attack, 800 benign) and approximately 2.3 million events. We captured system, network, and browser activity simultaneously on Windows endpoints. We labeled malicious events with ATT&CK technique IDs, covering 12 tactics and 53 techniques. We generated all attack data using real tools, including Remote Access Trojan (RAT), Command and Control (C2) tunnels, and cloud exfiltration. To demonstrate learnability, we fine-tuned three Small Language Models (SLMs) (Qwen2.5-1.5B, Llama-3.2-3B, Phi-4-Mini) using Low-Rank Adaptation (LoRA). We compared each against its base variant across ten metrics on two tasks: chunk classification and ATT&CK technique identification. Fine-tuning improved every model on every metric. Chunk classification accuracy rose from approximately 8% in the base variants to between 90% and 97% after fine-tuning. Technique identification remained challenging, with the best exact-match accuracy at 42%, although high partial-match scores show the models captured most of the underlying reasoning.
arXiv:2606.17074v1 Announce Type: new Abstract: Generative artificial intelligence (GenAI) is increasingly used for applications in the hardware and software domains. It purports to reduce the manual effort involved in the development and testing of complex systems before release. Within the hardware space, most tasks have focused on design automation of integrated circuits, particularly with hardware description languages. However, other types of hardware also exist! In this survey, we instead examine how GenAI has been and is being across the printed circuit board (PCB) design life cycle. This includes everything from supply chains, system specification, circuit design, layout and optimisation, validation and test, and PCB assembly and distribution. Through this lens we present a taxonomy of discovered works, categorising them according to their intent and contributions. This survey also identifies key technical challenges that GenAI faces in this space, such as domain-specific data scarcity and limited support for integration with existing PCB tools. Finally, future research directions are discussed: our survey shows that there are many opportunities remaining when considering how GenAI may be integrated into various tasks in PCB design and test.
arXiv:2606.17327v1 Announce Type: cross Abstract: The growing energy demand for computation is becoming increasingly unsustainable. Thermodynamic computing, which harnesses physical thermal fluctuations as a computational resource rather than suppressing them, offers orders-of-magnitude energy savings for probabilistic and combinatorial tasks. Pharmaceutical R&D, heavily reliant on computational optimization and sampling, is a natural application domain. Here we present what is, to our knowledge, the first concrete pharmaceutical application mapped to thermodynamic hardware with energy estimates grounded in prototype measurements. We reduce mRNA codon optimization, a combinatorial problem routinely solved in drug development, to sampling from an Ising model, making it directly executable on a thermodynamic sampling unit (TSU). Benchmarking three approaches (Potts sampling, Ising sampling, and a genetic algorithm baseline) on the SARS-CoV-2 spike protein, we find that all achieve comparable optimization quality (scores ~234-240), but energy estimates based on validated hardware models indicate that a TSU could solve this problem using approximately 10e6 times less energy than a conventional GPU. All code is released under an open-source license.
arXiv:2606.17902v1 Announce Type: cross Abstract: Magnetism has long been attributed to localized d, f, and even p electrons with strong correlations, whereas s electrons exemplified by hydrogen are reactive and tend to have their spins quenched, making s-electron-derived magnetism and long-range ordered magnetic crystals seem unattainable. Here we report a low-Z ferromagnetic crystal H13@(BN)12 using first-principles calculations, where thirteen hydrogen atoms are encapsulated within a (BN)12 cage and magnetism originates from the 1s electron of the central hydrogen atom. The crystal remains stable under ambient pressure owing to chemical precompression. Notably, the central hydrogen atom retains a magnetic moment of 1 {\mu}B, with long-range magnetic order established through multicenter bonding within the H13 aggregate and the intercell B-B network, while the zero orbital angular momentum of s electrons renders spin-orbit coupling (SOC) negligible as expected. Electronic structure analyses reveal that the large cavity and central negative electrostatic potential of the (BN)12 cage localize the hydrogen 1s electron, preventing spin quenching. Interestingly, under 16 GPa compression, the system transforms into a nonmagnetic metallic state driven by delocalized electrons of the central hydrogen atom. This study opens a pathway for constructing s-electron-driven magnetic materials and lays the foundation for developing low-energy consumption magnetic devices without SOC.
arXiv:2606.17077v1 Announce Type: new Abstract: Proton dissociation constants (pKa) are critical for functional molecule discovery and molecular modeling. Building on iBonD, the largest experimental pKa database established, we and other researchers have developed several methods including machine-learning-based empirical prediction and high-accuracy energy calculations. Despite this foundation, the rapid augmentation of high-quality pKa data remains fundamentally constrained. As part of this work, we performed large-scale regression-based pKa prediction on unlabeled molecular datasets using a collection of extensively optimized machine-learning models. The results indicate that, since the feature distributions of unlabeled molecular datasets, the pKa data distribution approximates normality, with extreme scarcity of tail-region samples. Although such augmentation is highly valuable for improving overall data availability and predictive modeling, it remains insufficient for efficiently discovering molecules with broad-spectrum pKa properties. To address this, we explore the targeted generation of molecules with sparse pKa properties from the vast chemical space. Given that traditional continuous latent space VAE-RNN methods for molecular generation suffer from insufficient stability and fail to demonstrate clear advantages in complementing sparse data, we design and implement a quantum-assisted sparse-pKa molecular generation. Feasibility is validated on a simulated quantum annealer, and superior extreme-value sampling is further achieved on physical coherent Ising machines (CIMs). (to be continued)
arXiv:2606.17082v1 Announce Type: new Abstract: End-to-end autonomous parking has emerged as a critical task within the realm of autonomous driving. However, existing methods suffer from black-box characteristics, lacking high-level semantic understanding and interpretability, which impedes the realization of seamless long-distance autonomous parking from the road to the target spot. To address these limitations, we propose ParkingTransformer, a novel framework that leverages multi-view perception and the scene understanding capability of Large Language Models (LLMs). By combining trajectory queries with LLMs implicit state features, our method interacts directly with historical information and raw sensor data to output planning trajectories, eliminating the need for dense Bird's-View (BEV) representations. To compensate for the inadequate spatial reasoning ability of LLMs, we introduce 3D positional encoding to explicitly inject spatial geometric awareness. Furthermore, a fixed-window streaming mechanism is designed for historical information processing, significantly improving long-term temporal processing efficiency and inference speed. Additionally, a coarse-to-fine decoding strategy is employed to progressively enhance trajectory precision. Extensive closed-loop experiments are conducted on the CARLA simulator and real-world vehicle platforms. The results demonstrate that our method achieves a driving score of 61.32 in CARLA simulator and an average success rate of 88.70% in real-world experiments, validating the feasibility and effectiveness of the proposed algorithms.
arXiv:2606.17092v1 Announce Type: new Abstract: Agentic systems are increasingly integrated with geographic information systems (GIS), where multi-agent coordination enables complex conversational and spatial analysis but introduces security risks. This work presents a security-oriented framework for risk identification, evaluation, and mitigation in a multi-agent GIS system while maintaining adaptability to broader agentic architectures. We test the agentic system of a commercial geospatial partner while developing a modular state-machine-based orchestration framework that abstracts agent behavior into reusable components. We evaluate robustness using a red-teaming framework with an adaptive attacker LLM and a deterministic judge that produces binary outcomes with supporting rationales across multi-turn attacks. We further improve resilience with a prompt optimization framework that treats prompts as structured signatures and injects adversarial demonstrations, enabling systematic security improvements without degrading task performance.
arXiv:2606.17100v1 Announce Type: new Abstract: Highly accurate stability boundary values for the ADER-DG method are obtained for arbitrary degrees $N$ of basis polynomials. In the linear case, stability is violated precisely when one of the matrix eigenvalues reaches $\lambda = -1$, regardless of the phase $\theta$. A rigorous mathematical framework for the stability is developed. The stability condition is significantly simplified, reducing it to the problem of calculating the roots of polynomials in the Courant number $\mathrm{CFL}$. The maximum of the Courant numbers $\mathrm{CFL}_{\rm max}(N)$ are calculated. These results are new and very convenient for practical use. A comparison of the obtained results with existing results reveals differences that may be significant for the selection of calculation parameters, especially for high degrees $N$. It is shown that widely used existing estimates $\mathrm{CFL}_{\rm max}(N) \propto 1/(2N+1)$ are overestimated. An interesting qualitative asymptotic $\mathrm{CFL}_{\rm max}(N) \propto (N+1)^{2}$ is obtained. A rigorous direct proof of the approximation is presented. Approximation orders $p = N+1$ for arbitrary degrees $N$ are rigorously derived. A set of numerical experiments is carried out to apply the ADER-DG method to solving both a linear advection equation and an Euler system of equations. The results obtained in these calculations confirm the theoretical results well. In particular, an excess of the Courant number over the $\mathrm{CFL}_{\rm max}(N)$ by even 1% in the linear case immediately leads to significant instability of the numerical solution. The obtained estimates of the boundary Courant number in the nonlinear case are somewhat underestimated -- by no more than 5%, which is due to the diffusivity and stability of the approximate Riemann solver. Empirical convergence orders are obtained, which are in good agreement with the theoretical results.
arXiv:2606.17101v1 Announce Type: new Abstract: This position paper examines a paradox encountered in UX research practice: a situation where real human participants delivered less authentic insights than AI personas might have, due to context-induced biases. We share our experience developing research-based AI personas using OpenAI's custom GPT builder and conducting a design thinking workshop with high-net-worth banking clients. The workshop setting, including a luxury hotel, present portfolio managers, and hospitality dynamics, introduced biases that compromised the feedback. We propose that AI personas offer an underexplored opportunity to mitigate certain human limitations in user research, and call for frameworks that help teams recognize when traditional research contexts introduce biases that AI personas might help avoid.
arXiv:2606.18009v1 Announce Type: new Abstract: The widely known Yin-Yang symbol (Taijitu) is based on nested circles of different radii whose areas are colored black and white such that the interface traces an $\mathcal{S}$-shaped curve. We address the question of how this symbol can be related to physical phenomena such as daytime and nighttime duration and the annual seasons. Using a simple dynamic model of daytime duration, we introduce the excess daytime fraction and reconstruct the symbol using the latitude of Zurich. In particular, we explain how the black and white areas are linked to the stability of Yin or Yang predominance. We further demonstrate that the Golden and Silver Ratios found in the geometry of the symbol carry meaning with respect to the Gregorian calendar. Finally, we construct an alternative Yin-Yang symbol using logarithmic spirals with the Golden Ratio as the growth parameter. The didactical quantitative derivation of the Yin-Yang symbol and its grounding in real-world observations can be regarded as a novel perspective on this iconic pattern.
arXiv:2606.17108v1 Announce Type: new Abstract: Supergranular flows organize transport in the solar photosphere over spatial and temporal scales much larger than granulation. While coherent vortical motions have been identified using objective Lagrangian diagnostics such as the Lagrangian-averaged vorticity deviation (LAVD), rotational coherence captures only one aspect of coherent flow organization. Here we introduce finite-time rotationally coherent sets (FTRCS) by combining the inflated dynamic Laplacian (IDL), which identifies finite-time quasi-material coherent regions, with LAVD-based rotational diagnostics. The IDL extracts coherent structures with finite lifetimes, while LAVD identifies those exhibiting enhanced intrinsic rotation. Application to photospheric velocity fields shows that instantaneous vortical features do not necessarily correspond to finite-time rotationally coherent structures. The analysis also illustrates the effect of compressibility: coherent sets may form through persistent contraction associated with convergent transport, rather than through the persistence of rotating material regions. The combined IDL--LAVD approach separates finite-time transport coherence from intrinsic rotational organization in time-dependent flows.
arXiv:2606.17645v1 Announce Type: new Abstract: Large language model (LLM) web agents are usually deployed as tool callers: each turn, the model reads a fresh page observation and emits one structured tool action. When every action is a low-level primitive, horizons grow quickly and so do policy-facing LLM completions, dominating latency and cost on benchmarks such as Mind2Web and WebArena. Recent systems therefore wrap repeated interaction fragments as web skills: callable tools built from successful trajectories or induced programs, so one call can replace several primitives. However, prior skill libraries are still triggered mainly by instruction similarity or coarse site metadata, which yields low skill reuse on held-out sites and leaves much of the potential step and token reduction on the table. We present SkillMigrator, an agent that learns reusable web skills and transfers them across sites by matching layout structure rather than specific element references. Each induced skill is stored as a transferable interaction pattern (TIP): the skill paired with a structural sketch of the snapshot at induction time. At test time, SkillMigrator retrieves TIPs by layout similarity and grounds their references on the live page. The rest of the stack is standard: accessibility-snapshot observations with stable references, and fixed tool calling over primitives plus skill invocations. Compared with the state-of-the-art approaches, SkillMigrator reduces the average LLM-action count on successful trajectories by 8-10% across both WebArena and Mind2Web at matched success rate.
arXiv:2606.17126v1 Announce Type: new Abstract: Singing style is a crucial aspect of a natural and expressive singing voice. Singers utilize singing styles to convey the feeling or emotion of the songs. Several works have been proposed to control singing style for making the more expressive singing voice. Recently, VibE-SVC successfully controls vibrato by predicting high-frequency F0 contour. In this paper, we introduce a singing voice conversion framework, called VibE-SVC2, to improve singing style conversion performance and controllability. The model offers control over two types of singing styles: a pitch style and a timbre style. For the pitch style, to resolve the pitch-energy entanglement issue that is unresolved in our previous work, we introduce a novel Energy Style Converter to address remaining style information in the energy contour. In addition, we propose a Zero-shot Pitch Style Converter, which mimics the pitch style of reference audio. To expand the controllability of the model, we propose vibrato rate scaling that is an independent control of vibrato extent, which is unavailable in VibE-SVC. For the timbre style, we extend the model to handle a variety of phonation styles. However, addressing specific styles such as vocal fry poses a challenge, as conventional F0 extraction often fails due to their inherent subharmonic characteristics, which degrades the conversion quality. To address this, we propose a novel Subharmonic Correction algorithm to refine the F0 contour for more natural timbre conversion. Through comprehensive objective and subjective evaluations, we demonstrate that VibE-SVC2 provides fine-grained, independent control over two types of singing styles, outperforming existing methods.
arXiv:2606.17162v1 Announce Type: new Abstract: Personalized presentation generation requires more than conditioning on a current prompt or template: agents must preserve stable user preferences across tasks, retain newly introduced preferences and constraints during multi-turn revision, and carry out local edits reliably. We propose MemSlides, a hierarchical memory framework for personalized presentation agents that separates long-term memory from working memory and further divides long-term memory into user profile memory and tool memory. User profile memory stores intent-conditioned profiles for round-0 personalization, working memory carries active preferences and session constraints across revision rounds, and tool memory stores reusable execution experience for reliable localized editing. MemSlides pairs this memory design with scoped slide-local revision, so targeted updates act on the smallest affected region instead of repeatedly regenerating the full deck. In controlled experiments, user profile memory improves persona-alignment judgments on a multi-persona, multi-intent profile bank, tool-memory injection improves closed-loop modify behavior in diagnostic matched-pair settings, and qualitative cases illustrate working memory's ability to carryover preferences. Taken together, these results suggest that effective personalization in presentation authoring depends on separating persistent user profiles, session-level working memory, and reusable execution experience across generation and localized revision.
arXiv:2606.17164v1 Announce Type: new Abstract: Prompting has become the primary interface between humans and generative AI, yet many natural language prompts remain fragile: roles, goals, constraints, and expected outputs are often buried in prose or left implicit. In agentic and software development workflows, a misread at the first handoff can propagate through every step, since a significant portion of agent failures stem from context ambiguities rather than model limitations. This paper introduces PromptMN, a pseudo-prompting domain-specific language that annotates natural language with compact, %-prefixed typed directives covering roles, goals, requirements, priorities, constraints, plans, inputs, and outputs. Semantic resolution lets authors write in any order while the model interprets directives by function. PromptMN sits between informal prompting and programming-style pseudocode: structured enough to be inspectable and reusable, yet lightweight enough for analysts, managers, developers, and stakeholders across the software development lifecycle (SDLC). PromptMN also pairs with reverse prompt engineering. Asking a model to restate a desired outcome as PromptMN lets users inspect the inferred roles, goals, constraints, and missing assumptions before acting, reducing repair cycles and yielding a reusable artifact for aligning people and AI tools. PromptMN's feasibility is evaluated across several frontier models, including Claude Fable 5, Claude Opus 4.8, Gemini 3.1 Pro, and GPT-5.5. The models correctly resolved PromptMN instructions, including complex structures such as repetition, conditionals, methods, and a prime-checking task, without fine-tuning. The same vocabulary applies across new codebases, maintenance, and redesign in the SDLC scenarios presented. While large-scale validation remains future work, these early results suggest PromptMN is a practical step toward clearer, more reviewable human-to-AI interaction.