Abstract

Corrosion creates a significant degradation mechanism in reinforced concrete (RC) structures, which would require a high cost of maintenance and repair in affected buildings. The right methods for structural retrofitting are, therefore, quite pivotal. Modification or rectification for any structure is normally categorized under three major categories: (i) Structural rehabilitation, (ii) Structural refurbishment, and (iii) Structural retrofit. All these activities have a thin but predominant line of differentiation. In contrast, the first two activities relate to enhancing the functional life of a structure that has been damaged under an adverse environment or condition. On the other hand, structural retrofit is an activity where the structure is updated in its basic parameters of sharing the load and supporting the superstructure. It improves the safety and durability of the structure. Further, when corrosion hits the structure, it results in a reduction in many important parameters to the extent that they go below their acceptable limits under national and international standards. It is indeed a perfect blend of structural dynamics or design engineering and functionality of the structure in the right proportion when retrofit action for a corrosion-hit structure is conceptualized and crafted. This paper aims to present and discuss various structural retrofitting methods for RC members affected by corrosion, supported by relevant real-world case studies.

Abstract

In stress-timed languages like English, word stress placement can change meaning, making automatic syllable stress
detection vital for Computer Assisted Language Learning
(CALL) systems, as L2 learners often misplace stress. Existing approaches rely on manually annotated boundaries or
forced alignment, which require extensive labeling and suffer
from alignment inaccuracies. We propose a fully boundaryindependent framework using convolutional neural networks
(CNNs) for hierarchical time-compression, converting framelevel embeddings into compact syllable-level representations
without explicit segmentation. To enforce the linguistic constraint that each word contains exactly one stressed syllable, we
also incorporate a Post-net2.0 loss. Experiments on the ISLE
corpus with German (GER) and Italian (ITA) learners show
consistent improvements over boundary-dependent and prior
boundary-independent SOTA models, achieving 94.86% (GER)
and 96.24% (ITA), with gains up to 18.67% and 16.12%.

Abstract

We consider replication-based distributed storage systems in which each node stores the same quantum of data and each data bit stored has the same replication factor across the nodes. Such systems are referred to as balanced distributed databases. When existing nodes leave or new nodes are added to this system, the balanced nature of the database is lost, either due to the reduction in the replication factor, or the non-uniformity of the storage at the nodes. This triggers a rebalancing algorithm, that exchanges data between the nodes so that the balance of the database is reinstated. The goal is then to design rebalancing schemes with minimal communication load. In a recent work by Krishnan et al., coded transmissions were used to rebalance a carefully designed distributed database from a node removal or addition. These coded rebalancing schemes have optimal communication load, however, require the file-size to be at least exponential in the system parameters. In this work, we consider a cyclic balanced database (where data is cyclically placed in the system nodes) and present coded rebalancing schemes for node removal and addition in such a database. Our algorithms have time-complexity polynomial in the system parameters. These databases (and the associated rebalancing schemes) require the file-size to be only cubic in the number of nodes in the system, which we argue is the minimum required for rebalancing on such databases. We bound the advantage of our node removal rebalancing scheme over the uncoded scheme, and show that our scheme has a smaller communication load. Further, we also argue that in some parameter regimes, cyclic databases have better asymptotic tradeoffs between the rebalancing load and the file-size, when compared to earlier work by Krishnan et al. In the node addition scenario, the rebalancing scheme presented is a simple uncoded scheme, which we show has optimal load. Finally, we show an information-theoretic lower bound for single node-removal rebalancing under the requirement of achieving a target cyclic database, and show that our achievable rebalancing loads are order-optimal (in the number of nodes), more specifically within a multiplicative gap of 2 from the lower bound obtained.

Abstract

Dynamic networks evolve as nodes, edges, and communities appear, disappear, or reorganize across time. Tracking this evolution while preserving the analyst's mental map poses a fundamental challenge: force-directed and static-community layouts reset spatial structure at each timeslice, breaking visual continuity. We present DyNetSpin, a visual analytics system that addresses this through two principled contributions. First, a History-Aware Spiral Layout extends community-centric spiral visualization [JK23] to the temporal domain via chronological accretion: communities are ranked by a weighted combination of logarithmic size (preventing large-community visual dominance) and historical barycenter (anchoring veteran communities near the spiral core across timeslices). Second, a Local Volatility metric classifies each node's directional temporal state-- Incoming, Outgoing, Transient, or Stable--within a three-timeslice sliding window, enabling analysts to distinguish structural growth from structural decline.

Abstract

Agentic kernel-synthesis loops are usually reported by their best number, which says little about whether the loop would survive a change of hardware or an honest verifier. We express GraphSynth, an LLM-driven Triton synthesis workflow for attention variants outside PyTorch's fused-kernel catalog, as a CHIA dataflow graph: four ChiaFunction stages, exactly one agentic edge, and a correctness gate that is never registered as a tool and therefore cannot be reached by the model that it judges. The loop replays end to end on a machine with no GPU and no credentials in about a minute, and runs for real on a GPU with only the language-model call bypassed. On an A100 the loop verifies all ten operators at 10.82× over eager PyTorch and 3.04× over torch.compile. Re-run on an NVIDIA L40S, which has 1.8× lower bandwidth and a ridge point 2.1× higher, it accepts 9 of 10 at 32.02×, against 11.76× for the same nine on the A100. The speedup nearly triples because the eager baseline is bandwidth-bound and the synthesized kernels are not: averaged over all ten operators the baseline is 1.795× slower on the L40S, against 1.800× predicted from bandwidth alone, while every synthesized kernel got faster. We also report what the first real execution exposed: three defects invisible to any replay, one of them the verifier being right while our own reference was wrong, and the one operator whose kernel does not fit Ada's shared memory.

Abstract

Waiting for work in latency-sensitive C++ systems involves a trade-off between tail latency, CPU consumption, and synchronization overhead. My work investigates three single-producer/single-consumer queue waiting strategies: mutex and condition-variable blocking, busy spinning, and C++20 std::atomic::wait.

A controlled experimental evaluation compares these strategies using matched preallocated ring-buffer storage under steady and bursty workloads. The primary study consists of 240 uninstrumented runs spanning 48 million measured events across four offered rates, two C++ standard-library implementations, and five repetitions per configuration. Per-event latency is decomposed into producer scheduling lateness and queue/service latency, with measurements of p99.9 tail latency and per-thread CPU consumption.

Results reveal that busy spinning maintains low tail latency at the cost of approximately one fully occupied CPU core. More significantly, identical atomic-waiting source code exhibits substantially different performance under libstdc++ and libc++. A separate Linux futex tracing study and inspection of the installed standard-library implementations identify different polling, backoff, and kernel-waiting mechanisms.

The findings demonstrate that synchronization performance depends not only on the C++ API but also on workload characteristics, measurement boundaries, CPU resources, and standard-library implementation.

Abstract

In this paper, we investigate the minimum obstacle displacement planning problem from a robot motion planning perspective.
The problem involves determining a feasible path to a goal location by displacing movable obstacles when no collision-free
path initially exists. We show that this problem is computationally challenging and, in particular, NP-hard when obstacles are
modeled as polygons in the plane. Besides an exact formulation of the minimum obstacle displacement problem generalizing
other problems in the literature, and the associated optimal solution, this paper proposes an approximate solution that is less
intensive from a computational standpoint, and differs from the optimal solution by a fraction of the optimal cost, being able
to trade off between path length and amount of obstacle displacements.

Abstract

Unreinforced masonry (URM) buildings constitute a substantial share of the global building stock
and remain highly vulnerable to seismic loading owing to their brittle response and negligible ten­
sile capacity. Flexural and shear failures are the most prevalent in-plane mechanisms governing
deformation capacity and collapse potential, making reliable identification of the governing failure
mode from crack patterns essential for damage assessment and retrofit decisions. However, inter­
pretation of cracking morphology remains inconsistent in practice, since it requires linkage to the
governing stress state and contextual parameters rather than appearance alone. This review syn­
thesizes the mechanics of flexural and shear crack formation in URM walls under in-plane loading,
as influenced by axial load ratio, aspect ratio, boundary conditions, and masonry anisotropy, and
examines transitional and mixed-mode responses that often cause post-earthquake misclassifica­
tion. A mechanics-based diagnostic framework is then proposed for crack-based identification of
failure mechanisms, operationalized as a sequential decision tree linking crack morphology, spatial
distribution, deformation kinematics, and governing parameters. A quantitative crack dominance
ratio (CDR) is proposed as a practical indicator for mixed-mode classification using experimentally
validated thresholds. The framework is illustrated against three experimental cases, producing clas­
sifications consistent with reported failure modes, and key research gaps are identified.

Abstract

Cross-modal scene and instance retrieval aims to retrieve a corresponding 3D environment or object given query 2D images, serving as a foundational mechanism for spatial reasoning in robotics and AR/VR. Existing methods achieve this by compressing the entire 3D map into a single global feature vector, which discards the fine-grained local geometrical structure in indoor scenes. In this work, we show that scene- and instance-level retrieval capabilities emerge from supervision on sparse, local 2D-3D correspondences, without any explicit global labels. Specifically, we propose $textbf{CrossWeave}$, a novel framework for cross-modal retrieval that learns sparse patch-point alignment, which alone outperforms global encoding methods. Additionally, an optimal transport-based aggregator is applied over the aligned features to attain further gain by clustering structurally redundant regions, freeing descriptor capacity for semantically distinct local features. CrossWeave achieves state-of-the-art performance on both scene- and instance-level retrieval on ScanNet and 3RScan with a $textit{single model}$, whereas existing methods require separate models for each task. Our results suggest that whole-scene encoding is not necessary for cross-modal retrieval, and that sparse local correspondence supervision is a more effective and flexible alternative.

Abstract

Gaussian Processes (GPs) are widely used as surrogates for black-box functions in sequential decision-making problems such as Bayesian optimization (BO), level set estimation (LSE), and Bayesian active learning (BAL). GP performance critically depends on kernels, and standard kernels can lead to suboptimal decisions under misspecification. To address this, we introduce HACK GPs (Hedge Adaptive Cumulative Kernels), a method that views kernel selection as an online learning with expert advice problem. HACK treats each candidate kernel as a GP "expert" and updates a distribution over experts online using AdaHedge, based on a loss received as a proxy for their ability to fit the function and align with the task objective. We provide two variants of HACK: (i) Mixture of Gaussians (MoG) and (ii) categorical sampling. We establish general guarantees showing that, under a loss-gap condition, the weight concentrates on the best kernel and the resulting acquisition function is close to that of the best expert. Empirically, we observe robust performance across BO, LSE, and BAL compared to standard kernels such as Squared Exponential and Matérn-5/2, as well as simple ensemble baselines