Papers

Self-evolving agents convert interaction feedback into persistent artifacts, such as memories or skills, which in turn guide subsequent decisions. As these artifacts are iteratively updated throughout an experience stream, the capabilities they support may evolve. Consequently, endpoint performance alone offers an incomplete view of self-evolution. Process-level evaluation is therefore essential to identify when a target capability emerges and whether later updates strengthen, preserve, or weaken it. Motivated by this, we propose EvoPathBench, a benchmark that tracks individual capabilities during artifact-level self-evolution. EvoPathBench fixes the base model, tools, freezes evolving artifacts at successive checkpoints, and evaluates the target capability on held-out episodes. This benchmark evaluates agent self-evolution using public trading data and calibrated trajectories. It tests three capabilities: generalization to unseen tasks, retention after unrelated learning, and rule adaptation to new evidence. Experimental results show that gains on similar unseen tasks often weaken under distribution shift, retention losses are concentrated in a minority of evolution paths, and no method achieves reliable rule adaptation. Moreover, while self-evolution enables agents to generate candidate artifacts with substantial held-out gains, the selected updates consistently fall short of realizing this potential. Together, these findings establish capability-level process evaluation as a foundation for analyzing self-evolution, identifying candidate evaluation and selection as key targets for improvement.

LLM-based agents increasingly operate in environments where they interact with users, tools, and external systems. Yet most security evaluations assume passive users and static control, ignoring the interactive dynamics that shape real agent behavior. We introduce DUMA-Bench, a benchmark and evaluation protocol for measuring agent security under dual-control interaction, where both the agent and the user can influence the shared environment state. DUMA-Bench extends $\tau^2$-bench ~\cite{barres2025tau} with adversarial environments covering eight vulnerability classes, including RAG poisoning, cross-agent manipulation, and unsafe output handling. We evaluate 14 models from five model families (OpenAI, Anthropic, DeepSeek, Qwen, and Z.ai) across eight domains and multiple user-behavior regimes. Across our experiments, introducing dual-control interaction increases the attack success rate from 26.9% to 41.1%. These results show that agent security is not solely a property of the model but emerges from the interaction between the model, the user, and the environment. DUMA-Bench provides a missing evaluation layer for studying security in realistic agent deployments.

Human demonstrations offer a scalable way to collect manipulation data, but their contacts may be unstable or infeasible when transferred to a robot hand. Collecting demonstrations directly on the target robot avoids this mismatch but substantially increases the cost of data collection. To address this trade-off, we present Touch2Robot, a framework that lets humans collect demonstrations while seeing how the target robot hand would contact the object. We capture human hand motion, tactile-glove measurements, and object motion during human manipulation. These recordings guide object-specific RL policies to reproduce the demonstrated object motion while favoring contacts consistent with the recorded human touch. We distill the learned behaviors into a unified real-time retargeter that maps incoming human observations and object geometry to robot hand configurations. During collection, the predicted robot configuration is synchronized with the tracked object pose in simulation to reconstruct robot-object contacts, which are visualized to help the demonstrator adapt subsequent interactions to the target hand. Across four real-world tasks, Touch2Robot improves average real-robot replay completion from 37.9% to 72.1% over visual-only feedback, while reducing the collection time per replay-successful demonstration from 58.6s to 18.2s. Reconstructed target-hand contacts achieve 44.2% F1 against real-robot tactile measurements, and policies trained on Touch2Robot demonstrations improve downstream Diffusion Policy performance by 29.1 percentage points over visual-only feedback. These results show that bringing robot touch into the human demonstration loop improves both the quality and efficiency of scalable dexterous data collection. Project webpage: https://Touch2Robot.github.io/.

InvestorNerd is a web-based platform (investornerd.org) designed to educate and democratize financial understanding by providing accessible, AI-powered investment and personal finance insights tailored to potential user profiles. The system addresses a key challenge in helping everyday individuals, especially those without formal financial education, make sense of investment options and personal financial decisions. The platform offers three interactive tools: Stock Insights - Allows users to input any Stock, Mutual Fund, or Exchange Traded Fund (ETF) ticker to receive a summary of relevant news sentiment and quantitative metrics. Stock Insights Questionnaire - Builds on stock insights by letting users specify a preferred risk tolerance and sector interest. It returns categorized investment tables based on volatility and sector, sortable by dividend yield, return percentages, and others. General Insights Questionnaire - Provides users with personalized insights based on answers to an age-income-expenditure-savings questionnaire. Outputs include possible actions regarding savings strategies, account types (e.g., Roth IRA, UTMA), and potential loan options (e.g., FHA). The output is designed to be clear, unbiased, and educational, empowering novice investors with practical insights. This paper details the design, implementation, and evaluation of InvestorNerd, demonstrating how generative AI and open financial data can be integrated to create scalable, insight-rich tools for financial literacy.

Language models can be adapted by changing the computations applied to individual tokens. Quantum circuits offer one such approach, but evaluating wider circuits inside a large model can be computationally demanding. Here we introduce HyperQ, which adds token-conditioned quantum residual branches to a frozen masked-diffusion language model. A quantum residual branch is a module in each transformer block that reads a token's hidden state, emits the coordinates of that token's circuit, executes it, and adds the measured values back through a residual connection. The backbone remains frozen, and only the added branches are trained. Within each branch, a lightweight circuit hypernetwork emits token-specific rotation angles, coupling strengths, and measurement axes in a shared sparse circuit structure. The required expectation values have an exact classical expression whose evaluation cost grows linearly with the qubit count, enabling circuits from 16 to 64 qubits to be trained within a 1.1-billion-parameter backbone. Across downstream benchmarks, increasing circuit width raises the average score from 47.65 to 54.30. At 64 qubits, HyperQ exceeds the backbone and its low-rank-adapted counterpart by 4.71 and 3.67 points, respectively. HyperQ is fine-tuned on 20,000 prompt-response pairs, compared with 200,000 for the classical baselines. These findings support token-conditioned circuit emission as a tractable architectural approach to quantum-augmented language modelling.

The fifth generation of mobile networks was standardised with an explicit mandate to close long-standing privacy and security gaps, mandating the concealment of the subscriber's permanent identity, resistance to generational downgrade, and protection against location tracking. Assessing whether these guarantees hold in operational networks, however, requires separating two sources of residual exposure that prior studies do not distinguish and do not evaluate in the wild: protocol-design limitations, which remain exploitable even against a fully specification-compliant deployment, and implementation gaps, which arise from incomplete or non-compliant implementations. We present 5G-Shark, a security assessment tool and methodology that turns a legitimate mobility procedure against the subscriber. Rather than relying on active jamming or malformed-packet injection, 5G-Shark manipulates the standardised cell-reselection criterion to pull a target User Equipment onto a self-created rogue cell, establishing an attack vantage with minimal service disruption. Then, the proposed methodology effectively performs the required interactions to expose the security risks of the system under test, classifying them into the aforementioned categories. Built solely from open-source stacks and Software Defined Radio hardware and evaluated against commercial 5G Standalone deployments, 5G-Shark requests subscriber identifiers, forces Radio Access Technology downgrade via crafted Registration Reject codes, and induces denial-of-service states. For each vector, we attribute the root cause to protocol design or deployment non-compliance. We further provide empirical evidence that in several commercial deployments, temporary identifiers are re-allocated in near-sequential steps that keep successive values linkable, a weakness that enables persistent user tracking despite correct subscriber ID concealment.

A mobile is a rooted full binary tree whose leaves carry positive integer weights. The imbalance of an internal node is the absolute difference between the total weights of its two child subtrees, and the cost of the mobile is the sum of these imbalances. In the unrestricted Balanced Mobiles problem, only the multiset of leaf weights is given: both the tree topology and the placement of the weights must be chosen so as to minimize the cost. The computational complexity of this unrestricted variant has remained open, although the variant with a prescribed topology is strongly NP-hard. We close this gap by proving that the decision version of unrestricted Balanced Mobiles is strongly NP-complete. Our reduction from Numerical 3-Dimensional Matching with Distinct Integers uses three widely separated numerical scales. Tight telescoping bounds force every threshold-achieving mobile into a canonical hierarchy, after which pairwise distinctness of the source integers collapses the hierarchy to single triples from which a valid numerical matching can be recovered.

Diffusion and flow models, as promising generative paradigms for speech enhancement, face a training–inference mismatch: training uses analytical path states, whereas inference recursively evaluates models on self-generated rollout states along discretized sampling trajectories. This mismatch causes prediction and discretization errors to accumulate. To address it, we introduce Corrective Forcing (CoF), a post-training paradigm that forces diffusion and flow models to learn from self-generated rollouts and correct their predictions. CoF corrects clean-speech predictions on rollout states toward the ground truth under dynamic sampling schedules, exposing the model to varying inference conditions. It further regularizes local evolution using locally corrected counterfactual transitions as references for factual transitions. By expressing model outputs through a shared clean-speech prediction parameterization, CoF applies the same post-training objective across diffusion and flow formulations. Experiments with SB-VE and OT-CFM demonstrate improvements in perceptual quality and reconstruction fidelity, together with robust performance across different numbers of sampling steps.

Readability assessment is essential for tailoring texts to intended audiences across educational, healthcare, and information retrieval domains. However, traditional readability formulas struggle to generalize across genres and languages, while supervised machine learning models rely on scarce, domain-specific annotated corpora, limiting their applicability–particularly for less-resourced languages. Large Language Models (LLMs) offer a highly scalable, multilingual alternative that requires no task-specific training, yet the impact of advanced prompting strategies on their performance remains underexplored. In this paper, we conduct a systematic benchmark of diverse open-source LLMs for multilingual readability assessment, focusing on the prediction of discrete readability levels required by educational frameworks. In addition to English, we evaluate our approach on a less-resourced language, Slovenian, to establish whether LLMs remain effective in low-resource settings. Specifically, we investigate the influence of explicit reasoning, demonstrating that Chain-of-Thought (CoT) prompting and reasoning-oriented models yield significant improvements over direct answering. Furthermore, our exploration of few-shot in-context learning reveals that providing just one labelled example per category (1-shot) substantially enhances prediction quality compared to zero-shot settings, with additional examples offering diminishing returns. By comprehensively comparing these approaches against traditional unsupervised metrics and state-of-the-art supervised baselines, we establish the viability of out-of-the-box LLMs as robust, cross-lingual readability assessors.

We compare the performance of the RPA and CHIRRUP decoding algorithms for second-order Reed–Muller codes. The RPA algorithm recovers a binary quadratic form associated with a skew-symmetric matrix $M$, whereas CHIRRUP recovers a $\mathbb{Z}_4$-valued quadratic form associated with a symmetric matrix $P$. We describe how the Gray map connects the evaluation vectors of these forms, inducing a rank-preserving correspondence between $M$ and $P$. We analyze how Euclidean distance between evaluation vectors is governed by matrix rank differences using Delsarte–Goethals sets $DG(m,r)$. We demonstrate that CHIRRUP uses repeated measurements more effectively than RPA on these ensembles. While a kernel-aware variant of RPA improves performance on singular quadratic forms by aggregating projection directions across cosets of $\ker M$, CHIRRUP's tree search combines multiple row estimates more efficiently on $DG(m,r)$ when $r < \frac{m-1}{2}$, outperforming vanilla RPA.

On-policy self-distillation fine-tuning (SDFT) learns new skills from demonstrations while reducing forgetting, but it always distils toward the full demonstration-conditioned teacher. This fixes teacher influence at the full-teacher endpoint, providing no control over how much demonstration information should be transferred at each prediction state. We introduce Information-Proximal SDFT (iSDFT), which instead treats the teacher as a budgeted source of information. At each token, iSDFT selects the distribution closest to the current student that satisfies a prescribed teacher-information constraint, yielding a closed-form exponential target with a locally determined tilt. To control cumulative drift, we further anchor the student to its frozen base policy. Across four heterogeneous LLM backbones and two specialisation tasks, iSDFT improves vanilla SDFT in 7 of 8 model-task settings and matches it in the remaining one. It also provides tighter retention on the original SDFT benchmark suite, with 73% of evaluations remaining within 0.5 points of the base model versus 52% for the strongest baseline, while achieving the largest mean improvement on all ten additional mathematics, coding, and competition-mathematics benchmarks. These results show that controlling how much and when teacher information is introduced improves specialisation while preserving broader capability.

As people turn to generative AI for financial advice, these systems can personalize how they communicate and what they say. Whether these forms of personalization shape decisions differently remains unclear. We conducted a preregistered 2 x 2 between-subjects factorial experiment (N=240): participants ranked three comparably viable stocks, discussed them with an AI, and reranked them. Participants perceived both forms of personalization, but only context-based personalization reliably changed ranking behavior: it increased reconsideration and moved rankings toward the AI's assigned recommendation. Participants felt more influenced without judging the AI as more correct, trustworthy, intelligent, likeable, or high-quality. Those initially farther from its recommendation moved more toward it while judging its advice less correct; exploratory analyses suggest greater susceptibility among lower-expertise participants. These findings show how personalized AI can steer decisions among defensible options with only a minimal evaluative trace, raising concerns for the design and governance of personalized decision support.

Power availability increasingly constrains the operation of AI inference fleets, creating a need for provisioning methods that jointly consider serving capacity and power consumption. Prefill–decode (PD) disaggregation has emerged as a prevalent architecture for large-scale inference serving. However, determining the appropriate numbers of prefill and decode instances is challenging because serving capacity depends jointly on workload characteristics, hardware constraints, queueing, and KV-cache reservations. Existing approaches largely rely on profiling and simulation, providing limited analytical insight into how provisioning decisions shape the tradeoff between serving capacity and power consumption. This paper develops an analytical framework for power-aware provisioning of PD-disaggregated AI inference. Given an inference workload and hardware, the framework models the serving capacity and average power consumption of a provisioned deployment. The serving-capacity model is derived from the joint distribution of input–output lengths and hardware compute and memory limits. In particular, it explicitly captures the coupling between prefill and decode induced by KV-cache reservations, as well as the impact of request queueing. On this basis, the power model determines per-instance power consumption as a function of normalized serving throughput. Together, the models determine the serving capacity–power Pareto front among candidate provisioned deployments, enabling the service provider to choose a provisioned deployment as the workload or available power changes.

Katz centrality is a well-established measure to identify and rank the most important nodes in complex networks by means of a linear system solve. Recent works have developed notions of Katz centrality for temporal, i.e., time-evolving networks. Their drawback is that small changes in the network structure may drastically change centralities across large parts of the network. Moreover, an unproportional effort in the network science community has been devoted to the study of node-based quantities while edge-based measures are far less explored. In this manuscript, we introduce a novel failure-aware temporal multiplex network model for computing edge-based Katz centralities for spatial networks. This class of networks admits the use of the line graph as network representation. We use a block-triangular and -banded supra-adjacency matrix representation, modeling inter-layer connections as low-rank matrices, which assigns a spatially constrained emphasis on up- or downdated edges. These could represent, e.g., breakages of water/gas pipes or obstructions in street networks in real-time infrastructure monitoring applications. We analyze the structure of the of the block matrix inverse with respect to entry decay in the blocks above the block-diagonal. This gives rise to two truncation approaches that allow drastic computational runtime reductions at the cost of introducing a controlled truncation error. Numerical experiments on a range of real-world spatio-temporal networks of size up to $2\cdot 10^8$ illustrate accuracy and efficiency with runtime gains of up to a factor of $679$.

When a language model reads an operation such as "Swap the contents of Box F and Box B", its forward pass writes keys and values for those tokens into the KV cache. Prior work on entity tracking establishes what models use: bindings are resolved at query time rather than stored as explicit latent state. We ask what they write at the operation span and how it is accessed. We split a forward pass into a frozen writer and a reader: the writer's cache is recomputed without gradients, while the reader sees only the instruction and operation tokens, with all state descriptions hidden, and is trained in isolation. Anything the reader recovers was therefore already present in the unmodified cache. On a synthetic boxes task, a base reader recovers $\leq 0.06$ of queried bindings against $0.75$–$1.00$ after training, and recoverability tracks the operation's read/write footprint. We find two modes of access. Across Llama-3.1-8B and Mistral-7B, operation-span transplants causally redirect which visible state is read even when the two worlds hold identical values, revealing a routing record. Isolation training preserves routing and adds direct access to the payload, the value the operation read, from the single operand-name token in a narrow mid-depth band (layers 12–15 of 32 in Llama-3.1-8B, 14–17 in Mistral-7B) — the same site that holds the routing record. The same recipe extends to further operations, ToMi and GSM8K, but is bounded by training coverage and costs open-book accuracy. Operation tokens thus leave localized, causally recoverable records that support both routing and direct payload access, though the model that writes them reads mainly the address they carry and not the value.

Nitrogen dioxide (NO2) is a common air pollutant, released into the atmosphere through the incomplete burning of fossil fuels, and associated with respiratory and cardiovascular diseases in humans. Ambient NO2 concentrations are regulated through air-quality limits assessed with a sparse network of fixed monitors. The revised EU Ambient Air Quality Directive (2024/2881) introduces a daily NO2 limit to be met from 2030. At present, neither the regulatory monitoring network nor existing coarse, annual-mean models can resolve NO2 concentrations at the spatio-temporal resolutions necessary to assess compliance. Here we map NO2 across Europe at hourly and 10m resolution with a machine-learning model that combines ground monitors with satellite, reanalysis, land-use, traffic and emission data and returns a calibrated predictive distribution at every location. Validated against held-out regulatory monitors and independent citizen-science campaigns, the maps resolve high-resolution spatiotemporal NO2 gradients for 110 metropolitan areas in Europe. We reconstruct the daily compliance statistic across those regions and find limit breaches in 91 EU air quality zones deemed compliant by the regulatory monitoring network, covering a population of approximately 135M. Beyond air quality zones and monitor locations, an estimated 9-9.4% (20M) of the population in mapped regions lives in areas where the daily NO2 limit is breached. The high-resolution maps offer a route to population-scale assessment of compliance with the 2030 limits.

Robot manipulators are monitored by constraint-specific indicators whose units and gradient scales are not comparable, so they do not provide a common measure of the configuration-space motion remaining before violation. We define the feasibility distance field (FDF) as the distance, under a fixed positive-definite joint-space metric, to the union of infeasible configuration sets. Classical distance-to-set theory gives 1-Lipschitz continuity, almost-everywhere differentiability, and unit dual-gradient norm wherever the nearest projection is unique. The robotics contribution is an admissibility analysis showing when practical constraints define non-empty closed sets. We derive admissible formulations for external and self-collision, joint limits, dexterity, Cartesian and task-projected compliance, joint torque under payload, and dynamic manipulability. Since every field uses the same metric, heterogeneous constraints compose by a pointwise minimum, conditioned constraints retain a fixed distance space, and multi-robot constraints produce block-sparse gradients that identify which robots must react. We generate projection-based labels and train neural approximations with a distance loss and an Eikonal penalty. Simulations on a UR5e and a dual-arm cell evaluate seven fields using value, projection, sign, gradient, composition, and moving-obstacle diagnostics. Across 8,000 configurations, the largest feasible-side secant ratio is 0.920, mean learned gradient norms range from 0.994 to 0.998, and projection residuals range from 0.011 to 0.034 rad. Across 24 random obstacle paths, the external and composed collision fields achieve 90.4% and 91.6% success within 3 cm, with sign-error rates below 2%. The results support a common configuration-space margin and identify approximation errors near medial axes and sparsely sampled boundaries.

Autonomous parking in nonconvex and narrow environments remains challenging. Although optimal-control methods can explicitly enforce vehicle dynamics and collision constraints, nonconvexity compromises solver robustness and can cause failures. Large language models (LLMs) exhibit strong semantic reasoning capabilities, but directly generating dense trajectories makes it difficult to guarantee physical feasibility. We introduce SE-LLM-OCP, a unified framework in which LLMs make high-level discrete maneuver decisions, while an optimal-control module enforces low-level vehicle dynamics and collision constraints. Online, the LLM proposes sparse maneuver plans, decomposing the parking task into a sequence of short-horizon trajectory-optimization problems. A low-level solver then sequentially solves optimal-control problems. If the solver fails, the LLM aggregates failure evidence from the solver and validation stages to guide replanning. Offline, SE-LLM-OCP automatically evolves a structured decision-making knowledge base from scratch, driven by accumulated online failures. We validate our proposed framework in simulation on a car-like vehicle model and on a differential-drive robot. Our experimental results show that SE-LLM-OCP enables safer autonomous parking in narrow scenarios and demonstrates transfer of the same maneuver representation to a different kinematic platform.

A/B testing requires large sample sizes, long timelines, and significant costs. When auxiliary predictions of experimental outcomes are available from machine learning models, uncertain prediction quality precludes replacing human experiments entirely, yet these predictions may still contain useful signal. We propose a principled framework for learning-augmented hypothesis testing that leverages predictions of unknown quality to reduce sample sizes while maintaining statistical validity. Predictions naturally vary in granularity, from coarse aggregate signals to fine-grained individual-level estimates, and our framework addresses both ends of this spectrum: (1) for population-level directional predictions, where only a binary signal on the treatment effect sign is available, we use an asymmetric test and prove consistency and robustness bounds within the learning-augmented algorithms paradigm; (2) for individual-level predictions, we introduce Generalized PPI++ (GPPI), extending Prediction-Powered Inference to handle nonlinear prediction errors through higher-dimensional transformations. Both methods benefit from accurate predictions while remaining robust to inaccurate or adversarial ones. We validate our framework using persona-based LLM simulations, where AI agents equipped with user personas predict individual behavior, as a natural prediction source spanning both granularity levels. Experiments on four real-world datasets demonstrate that our methods, combined with persona-based predictions, substantially reduce experimental costs while preserving rigorous statistical validity.

The High-Luminosity LHC (HL-LHC) will demand order-of-magnitude gains in analysis throughput, and increasingly those gains must come from GPUs that are not made by a single vendor. Leadership-class systems such as El Capitan, Frontier and LUMI are built on AMD accelerators, yet the Scikit-HEP analysis stack—and Awkward Array in particular—has grown up CUDA-first. We report on $rawkward$, a Rust-backed kernel engine that adds a ROCm/HIP backend for Awkward Array's nested, jagged, variable-length data structures. Our central finding is that a naive source-level port of CUDA kernels to HIP loses $5$–$10\times$ in performance on irregular kernels, because AMD's $64$-lane wavefronts, higher register pressure and more expensive divergence behave fundamentally differently from NVIDIA's $32$-thread warps. We show that a small, reusable set of optimization patterns—loop flattening, $128$-bit vectorized loads, splitting fused kernels, and profile-guided launch configuration—recovers CUDA-class performance without changing the public API. A Rust macro-and-match dispatch layer keeps a single, backend-agnostic call site while emitting vendor-specific kernel strategies, and the type system enforces buffer-size and lifetime correctness at compile time. On a two-socket AMD Instinct MI210 node we measure GPU speedups from $1.03\times$ (bandwidth-bound $sum$) up to $12.5\times$ ($count$) over $128$ EPYC~7763 CPU cores, and the Rust CPU kernels match or beat on aggregate the incumbent C++ kernels (geometric-mean runtime ratio $0.37\times$ across twelve kernels). We argue that these patterns constitute a practical recipe for performance-portable, vendor-agnostic HEP analysis kernels.

Multi-organ segmentation using deep learning requires large amounts of annotated patient data; however, institutions often lack sufficiently large and diverse annotated datasets. Privacy constraints further prevent institutions from sharing patient data to overcome this limitation. Moreover, due to the labor-intensive nature of annotation and the scarcity of diverse expertise, institutions typically have labels for only a small portion of their local data, leaving the larger unlabeled portion unused. In this work, we propose a flexible semi-supervised federated multi-task student-teacher framework that leverages federated learning (FL) to improve multi-organ segmentation using both labeled and unlabeled data across participating sites. At each communication round, the proposed framework initiates local training, where clients with labels for the same task form a federation to produce an aggregated teacher model. The resulting teachers generate task-specific features for all data at each client. Subsequently, all clients form a second federation to train a multi-task student model with a shared encoder and task-specific decoders that replicate the teacher-generated features across all segmentation tasks. The aggregated student model is then used to update the local teachers and initiate the next training round. Extensive experiments demonstrated the effectiveness of the proposed method compared with local and federated single-organ models, yielding an average performance gain of 13 percent across clients. The experiments also demonstrated the impact of multi-task learning and unlabeled data and the applicability of the framework in relaxing labeled-data requirements for client participation. The code is available at https://github.com/AshknMrd/FedMust.

Latent world models learn predictive representations for autonomous driving, but the relational semantics these states preserve often remain implicit. We investigate whether traffic scene graphs can serve as privileged semantic supervision for latent world representations. Building on LAW, we construct actor-centric scene graphs from nuScenes 3D annotations, encode their serialized relational structure using a frozen text embedding model, and align the visual latent representations with this semantic target during training. We remove the supervision branch at inference, so it requires neither scene graphs nor 3D annotations and adds no test-time computation. On nuScenes, our method reduces average trajectory L2 error from 0.661 to 0.622 (5.9%) and collision rate from 0.456 to 0.217 (52.4%) relative to our retrained LAW baseline. It also outperforms an unstructured caption-style semantic target, supporting the benefit of explicit relational structure for latent world-model representation learning.

How should natural language processing models be selected and adapted for global health literature in environments where annotated data and computational resources are limited? This thesis investigates these challenges through experiments on semantic tag discovery, named entity recognition (NER), and multi-label topic classification. First, skip-gram word2vec models trained on progressively larger specialized corpora are compared with BioWordVec to assess how corpus size and domain context influence tag discovery. Vocabulary coverage and qualitative evaluation indicate that broader coverage does not necessarily yield more useful domain-specific associations. The analysis then turns to entity extraction, comparing convolutional spaCy models with a RoBERTa-based transformer on 1,000 annotated sentences. Under a lenient scoring protocol, the transformer achieves 0.80 micro-F1 versus 0.65-0.69 for convolutional models, but takes 82 seconds rather than 5-6 seconds. This trade-off motivates fine-tuning convolutional models and integrating a disease recognizer that achieves 81.33% test F1 on the NCBI Disease Corpus. Combined with PDF preprocessing, entity filtering, and MeSH enrichment, the resulting pipeline supports document-level indexing. To complement entity extraction with thematic annotation, MiniLM-based few-shot classification is compared with BART-MNLI zero-shot inference across 50 topics and 1,000 handcrafted test sentences. BART-MNLI achieves 95.2% single-label accuracy versus 59%; reported multi-label accuracies are 88% and 32% under partly manual assessment. However, its higher inference cost limits practical integration. The results show where domain specialization and lightweight adaptation offer practical value, and where transformer accuracy justifies higher inference costs, providing an empirical basis for building knowledge systems under resource constraints.

In the vertex cover interdiction problem, we are given an undirected graph $G=(V,E)$, two integers $t$ and $k$ and a vertex subset $B\subseteq V$, and we are asked to find a set $X \subseteq B$ with $|X|\leq t$ such that $X$ hits (i.e., intersects) all the vertex covers of $G$ of size at most $k$. Recently, Grüne and Wulf proved that the problem is $\Sigma_2^p$-complete. However, their reduction relied on the fact that the vertex cover problem is NP-complete. This, in turn, means that we do not know the complexity status of the vertex cover interdiction problem when the input graph is restricted to a bipartite graph since the vertex cover problem can be solved in polynomial time for bipartite graphs. One of our main results shows that the vertex cover interdiction problem is NP-complete for bipartite graphs. In contrast, when $k$ is restricted to the minimum vertex cover size, i.e., we are only required to hit all the minimum vertex covers, we show that the vertex cover interdiction problem can be solved in polynomial time for bipartite graphs. This motivates us to study the parameterized complexity of the vertex cover interdiction problem for bipartite graphs when the difference of $k$ and the minimum vertex cover size is taken as a parameter. With this parameter, we show that the problem is $\mathrm{W}[1]$-hard, but can be solved in polynomial time when the parameter is constant (i.e., in XP time). We also show that the problem is fixed-parameter tractable when parameterized by $k$.

We establish arbitrary-order a priori regularity estimates for the reverse heat flow in the Ornstein–Uhlenbeck setting, that is, for the probability-flow ODE of score-based diffusion models. For compactly supported, possibly singular targets satisfying a dimension-uniform doubling condition on supporting-cap masses, we bound the iterated material derivatives of the flow and their first spatial derivatives pointwise, with constants independent of the dimension $d$ and explicit in time and support radius. The proof rests on two ingredients: a spatially uniform bound on normal-direction posterior fluctuations, which is the only point where the geometry of the target enters, and a centered posterior-moment normal form closed under material differentiation. This algebraic structure, together with a weighted regularity calculus, yields quantitative bounds at every order, including the mixed space–time directional derivatives required by Runge–Kutta stages. As an application, for exact-score sampling stopped at forward time $\delta>0$, order-$p$ Taylor and explicit Runge–Kutta schemes reach accuracy $\varepsilon$ in $\widetilde O(d^{1/p}\varepsilon^{-1/p})$ steps in total variation, and the Taylor scheme in $\widetilde O(d^{1/(2p)}\varepsilon^{-1/p})$ steps in $W_2$.

Tactile sensing is increasingly being incorporated into learning-based robotic manipulation, yet many existing approaches rely on spatially distributed sensors. Here we introduce {SpectRobot}, a framework that transforms single-point tactile signals into compact time-frequency spectrograms. These spectrograms encode high-bandwidth tactile histories as fixed-size image-like representations. They can be processed by standard vision encoders and integrated into learning pipelines originally developed for vision, while preserving temporal and frequency information unavailable to conventional cameras. Rather than increasing spatial density through arrays of tactile elements, SpectRobot exploits the rich dynamics contained in sparse, high-bandwidth single-point measurements. In our implementation, the sensors are mounted away from the contact surface while remaining mechanically coupled to it, reducing direct exposure to wear and potentially improving robustness in harsh environments and for long-term deployment on dexterous robots. Our experiments demonstrate that: (1) a robot can exploit single-point vibration signals to solve a visually occluded manipulation task; (2) temporal history strongly influences policy performance, while sensing bandwidth controls the spectral information available, with measurements extending to 100~kHz; and (3) the same representation can be used across different tactile sensing technologies mediated by acceleration, force, or strain. We further show that capabilities previously associated with research-grade instrumentation can be accessed using readily available, off-the-shelf hardware. We believe that broader access to high-bandwidth tactile sensing could facilitate the integration of contact dynamics into embodied learning systems and, for some tasks, offer an alternative or complement to increasing the spatial density of tactile sensing.

Answering epidemiological questions from real-world clinical data requires medical coding, schema-aware SQL, and validation of implicit choices about populations, denominators, and time. We present Ascent, an agentic system that exposes medical coding, question answering, and cohort analysis through a shared Model Context Protocol tool surface for standardized and native schemas. We introduce EpiTrap, a dataset testing whether systems avoid recognized pharmacoepidemiological errors, and compare a fixed pipeline with agents across models and orchestrators. With capable models, agents improve accuracy over the fixed pipeline by an average of 27 and 20 percentage points on native and standardized schemas, respectively. These gains require more tool calls and longer runtimes. Experience from real projects highlights the system's value for feasibility assessment, diagnostic iteration, and expert-guided analysis.

The proposed pipeline integrates a representation-flow convolutional neural network with a dynamics- and uncertainty-aware tree-based Gaussian Process classifier. In this framework, latent motion dynamics are exploited both as discriminative representations and as a source of input uncertainty, enhancing robustness against temporal variations and abnormal motion transitions. Compared with conventional deep learning approaches, the proposed strategy requires less training data and offers improved computational efficiency. To further improve classification performance, we introduce novel semantic-aware compound kernels that effectively capture semantic, flow, and dynamic information embedded in surgical video features. In addition, uncertainty-aware kernels are developed to strengthen the robustness and practical applicability of the compound kernel framework. The proposed method is evaluated on two benchmark datasets, namely the JIGSAWS and the Cataract-LMM (Capsulorhexis) datasets. Experimental results demonstrate strong performance across both datasets, including the LOSO and LOUO evaluation protocols on JIGSAWS, including the subject-independent LOUO protocol on JIGSAWS, on which the framework attains a mean accuracy of \ph{96.9}%; results under the within-subject LOSO protocol are reported for comparability with prior work, achieving competitive accuracy while substantially reducing computational cost. Overall, the proposed pipeline provides an efficient and accurate framework for video-based surgical skill assessment.

AI coding interfaces present generated code either all at once or token-by-token. These rendering strategies reflect model generation rather than how programmers actually read code: selectively, non-linearly, and guided by the structure. We argue that code rendering is a first-class interaction primitive that shapes how programmers read and understand code. To explore this design space, we introduce structured rendering, a technique that reveals code in semantically meaningful chunks derived from its syntactic hierarchy, exposing high-level structure before low-level details. To isolate rendering effects on visual attention, we conducted an eye-tracking study with 53 participants comparing static, character-based, and structured rendering. Our findings show that rendering alters visual attention and reading behavior: dynamic rendering induces fewer but longer fixations and more sustained focus, while structured rendering further guides attention toward semantically meaningful units and supports high-level understanding. We release the anonymized dataset with an interactive demo (https://codegaze.vercel.app/) to support future research.

Prior case retrieval (PCR) aims to identify precedent cases relevant to a given query case. Existing PCR benchmarks and methods predominantly operate at the document level, treating entire judgments as the unit of relevance. This formulation is suboptimal for legal practitioners, as judgments address multiple legal issues and only a small subset of paragraphs is relevant to a particular query. Addressing this gap, we introduce UK-PRBench, a benchmark for paragraph-level precedent retrieval in UK case law, constructed from judgments obtained from the UK National Archives and covering a broad range of UK courts and tribunals. Furthermore, we evaluate state-of-the-art retrieval models and establish baseline results. Our experiments show that paragraph-level precedent retrieval remains challenging for current retrieval approaches, highlighting substantial room for improvement. UK-PRBench provides a standardised benchmark for evaluating fine-grained precedent retrieval and advancing retrieval systems for the UK legal domain.