arXiv subjects

Neil Rodrigues

Publications and source records attributed to Neil Rodrigues.

7 recordsLinked to original sources

arXiv2026

Training-Free Inference Acceleration for Video Large Language Models: Methods, Trade-offs, and Evaluation

Video Large Language Models (VLMs) have demonstrated strong capabilities in video understanding and reasoning, but processing long visual sequences incurs substantial computational, latency, and memory overhead. This survey examines training-free inference acceleration, focusing on reducing inference costs while preserving task-relevant spatiotemporal evidence. We organize existing methods by their intervention stages in the inference pipeline, covering input, visual encoding, projector, prefill, and decoding. Post-training quantization (PTQ) is discussed separately as a cross-stage extension with varying calibration requirements. We analyze how intervention position affects evidence preservation and computational savings, and why local efficiency improvements may not translate into end-to-end acceleration. We complement this analysis with quantitative evidence from prior studies and our own experiments under a shared evaluation setting to assess accuracy-latency trade-offs.

cs.CV↗

arXiv2026

OpenHome: From Perception to Insight in Smart-Home Camera Videos

Vision language models (VLMs) have been increasingly adopted in consumer smart-home camera systems to interpret recorded videos and provide insights. However, video clips captured from smart-home cameras are different from curated web videos dominating existing VLM benchmarks. Smart-home video clips often do not contain subtitles or narration and are captured from different viewing angles, under various lighting conditions, and in diverse scenarios. However, the capabilities of the model in this domain, under sparse visual signals and limited context, are not fully assessed by existing VLM benchmarks. To close this gap, we introduce OpenHome, a visual-only, four-option multiple-choice benchmark for evaluating smart-home video understanding. It contains 2,254 questions on 1,118 real-world video clips and organizes 10 tasks into three capability levels. Perception evaluates visible content, Reasoning evaluates causal, spatial, and temporal relationships, and Insight evaluates subject intent, supporting evidence and appropriate homeowner responses. Based on the dataset, we describe the task design, data pipeline and evaluation protocol, and report results for 32 closed-source and open-weight VLMs. The best model, Gemini 4 Argon, reaches 81.2% overall accuracy. Frontier closed-source models outperform open-weight models on every level. The best closed-source model leads the best open-weight model by 16.3% on Perception, 22.4% on Reasoning and 37.4% on Insight.

cs.CV↗

arXiv2026

TeQCo: Temporal Quota Consistency for Training-Free Video Token Compression

Video token compression must control both temporal coverage and task relevance, but sequential stages can conflict: a content-driven allocation stage may be overwritten by a global ranking that removes entire temporal positions. We identify this as an allocation conflict and propose TeQCo, a training-free framework that resolves it through temporal quota consistency. TeQCo allocates candidate capacity from first-order changes in normalized frame representations and aggregates all tokens within each frame into a compact candidate set. Candidates are then scored by prior-corrected task attention and selected within integer frame quotas inherited from the first-stage counts. Content variation determines temporal capacity; question relevance determines which candidates fill it. This design guarantees an exact token budget and a per-frame coverage floor whenever feasible. On the 2,700-question Video-MME benchmark with Qwen3-VL-8B, TeQCo achieves 68.52% accuracy at 510 ms TTFT, a 32.9% TTFT reduction with a 1.00-percentage-point accuracy decrease.

eess.IV↗

arXiv2026

ContextTom: Training-Free Video VLM Acceleration with Early Spatial Context

Video vision-language models process thousands of visual tokens before producing an answer. Reducing these tokens early saves computation, but also determines which evidence reaches the language model. We introduce ContextTom, a training-free framework that separates the timing of temporal and spatial compression. Its central idea is to compress time before vision encoding while preserving spatial coverage through early multimodal processing. A fixed-budget temporal partition produces representative medoid anchors, supplemented with bounded local detail from neighboring observations. The anchors retain their complete spatial grids through a short decoder prefix, allowing question tokens to access the visual evidence and completing Qwen3-VL's multi-level visual injections. ContextTom then selects spatial representatives using query attention and agreement across visual feature levels, and continues directly from their computed states. This schedule reduces vision computation and later decoder workload while reusing the prefix. On Qwen3-VL-8B, ContextTom achieves 76.22% accuracy on Video-MME Short with a mean time-to-first-token (TTFT) of 123.5 ms, yielding a 2.77x speedup over full-input inference. It matches the full-input accuracy of 72.20% on EgoSchema with a 2.82x speedup and achieves a 5.54x speedup on LongVideoBench at 59.24% accuracy.

eess.IV↗

arXiv2025

On interdependence of instabilities and average drop sizes in bag breakup

A drop exposed to cross flow of air experiences sudden accelerations which deform it rapidly ultimately proceeding to disintegrate it into smaller fragments. In this work, we examine the breakup of a drop as a bag film with a bounding rim resulting from acceleration induced Rayleigh-Taylor instabilities and characterized through the Weber number, \textit{We}, representative of the competition between the disruptive aerodynamic force imparting acceleration and the restorative surface tension force. Our analysis reveals a previously overlooked parabolic dependence ($\sim We^2$) of the combination of dimensionless instability wavelengths $({\barλ}_{bag}^2/ {\barλ}_{rim}^4 {\barλ}_{film})$ developing on different segments of the deforming drop. Further, we extend these findings to deduce the dependence of the average dimensionless drop sizes for the rim, $\langle{\bar{D}}_{rim}\rangle$ and bag film, $\langle{\bar{D}}_{film}\rangle$ individually, on $We$ and see them to decrease linearly for the rim ($\sim We^{-1}$) and quadratically for the bag film ($\sim We^{-2}$). The reported work is expected to have far-reaching implications as it provides unique insights on destabilization and disintegration mechanisms based on theoretical scaling arguments involving the commonly encountered canonical geometries of a toroidal rim and a curved liquid film.

physics.flu-dyn↗

arXiv2020

Mine Tunnel Exploration using Multiple Quadrupedal Robots

Robotic exploration of underground environments is a particularly challenging problem due to communication, endurance, and traversability constraints which necessitate high degrees of autonomy and agility. These challenges are further exacerbated by the need to minimize human intervention for practical applications. While legged robots have the ability to traverse extremely challenging terrain, they also engender new challenges for planning, estimation, and control. In this work, we describe a fully autonomous system for multi-robot mine exploration and mapping using legged quadrupeds, as well as a distributed database mesh networking system for reporting data. In addition, we show results from the DARPA Subterranean Challenge (SubT) Tunnel Circuit demonstrating localization of artifacts after traversals of hundreds of meters. These experiments describe fully autonomous exploration of an unknown Global Navigation Satellite System (GNSS)-denied environment undertaken by legged robots.

cs.RO↗

arXiv2019

PST900: RGB-Thermal Calibration, Dataset and Segmentation Network

In this work we propose long wave infrared (LWIR) imagery as a viable supporting modality for semantic segmentation using learning-based techniques. We first address the problem of RGB-thermal camera calibration by proposing a passive calibration target and procedure that is both portable and easy to use. Second, we present PST900, a dataset of 894 synchronized and calibrated RGB and Thermal image pairs with per pixel human annotations across four distinct classes from the DARPA Subterranean Challenge. Lastly, we propose a CNN architecture for fast semantic segmentation that combines both RGB and Thermal imagery in a way that leverages RGB imagery independently. We compare our method against the state-of-the-art and show that our method outperforms them in our dataset.

cs.CV↗