A stencil computation repeatedly updates each point of a d-dimensional grid as function itself and its near neighbors. Parallel cache-efficient algorithms based on "trapezoidal decompositions" are known, but most programmers find them difficult to write. The Pochoir compiler allows programmer write simple specification in domain-specific language embedded C++ which the then translates into high-performing Cilk code that employs an efficient parallel cache-oblivious algorithm. supports...

With the rapid development of wireless communication and flight control technologies, unmanned aerial vehicles (UAVs) have been widely used in multiple application scenarios. A typical scenario is massive crowd management multi-millions annual Hajj Pilgrimage to Mecca where UAVs are utilized conduct monitoring by carrying sensory devices. The safe a UAV crucial for ensuring successful execution missions. aim overcome disadvantage caused ground station intrusion detection, combination...

State-of-the-art cache-oblivious parallel algorithms for dynamic programming (DP) problems usually guarantee asymptotically optimal cache performance without any tuning of parameters, but they often fail to exploit the theoretically best parallelism at same time. While these achieve cache-optimality through use a recursive divide-and-conquer (DAC) strategy, scheduling tasks granularity task dependency introduces artificial dependencies in addition those arising from defining recurrence...

With the large-scale commercialization of 5G, global industry has started exploration next generation mobile communication technology (6G). From Internet, to IoT, and then smart connection everything, 6G will transform from 5G's service objects people things intelligent networking agent that supports human–machine–object. networks should have characteristics ubiquitous intelligence perception, which poses challenges for network construction. Therefore, we propose a Semantic Communication...

We present AUTOGEN---an algorithm that for a wide class of dynamic programming (DP) problems automatically discovers highly efficient cache-oblivious parallel recursive divide-and-conquer algorithms from inefficient iterative descriptions DP recurrences. AUTOGEN analyzes the set table locations accessed by when run on small size, and identifies access pattern corresponding provably correct solving recurrence. use to autodiscover several well-known problems. Our experimental results show...

The nested parallel (a.k.a. fork-join) model is widely used for writing programs. However, the two composition constructs, i.e. "||" (parallel) and ";" (serial), that comprise nested-parallel are insufficient in expressing "partial dependencies" a program. We propose new dataflow construct "↝" to express partial dependencies algorithms processor- cache-oblivious way, thus extending Nested Parallel (NP) Dataflow (ND) model. redesign several divide-and-conquer ranging from dense linear algebra...

Iterative wavefront algorithms for evaluating dynamic programming recurrences exploit optimal parallelism but show poor cache performance. Tiled-iterative achieve complexity and high are cache-aware hence not portable cache-adaptive. On the other hand, standard cache-oblivious recursive divide-and-conquer have serial often low due to artificial dependencies among subtasks. Recently, we introduced (COW) algorithms, which do any dependencies, they too complicated develop, analyze, implement,...

The advantages of intelligent fabrics enable health service systems more efficient and safe to perceive transmit, where massive sensitive data related behavior are collected through the multi-source flexible sensor network. However, due its complex network structure, face challenge low-power transmission processing in fabric space. Recently, Deep learning-powered semantic communication (DeepSC) has emerged as a promising paradigm. To tackle above problem, we introduce concept cognition into...

Self-supervised learning has achieved great success in both natural language processing and 2D vision, where masked modeling is a quite popular pre-training scheme. However, extending masking to 3D point cloud understanding that combines local global features poses new challenge. In our work, we present Point-LGMask, novel method embed contexts with multi-ratio masking, which effective for self-supervised feature of clouds but unfortunately ignored by existing works. Specifically, avoid...

Existing Transformer-based models for point cloud analysis suffer from quadratic complexity, leading to compromised resolution and information loss. In contrast, the newly proposed Mamba model, based on state space (SSM), outperforms Transformer in multiple areas with only linear complexity. However, straightforward adoption of does not achieve satisfactory performance tasks. this work, we present Mamba3D, a model tailored learning enhance local feature extraction, achieving superior...

We present A utogen —an algorithm that for a wide class of dynamic programming (DP) problems automatically discovers highly efficient cache-oblivious parallel recursive divide-and-conquer algorithms from inefficient iterative descriptions DP recurrences. analyzes the set table locations accessed by when run on small size and identifies access pattern corresponding provably correct solving recurrence. use to autodiscover several well-known problems. Our experimental results show...

The state-of-the-art "trapezoidal decomposition algorithm" for stencil computations on modern multicore machines use recursive divide-and-conquer (DAC) to achieve asymptotically optimal cache complexity cache-obliviously. But the same DAC approach restricts parallelism by introducing artificial dependencies among subtasks in addition those arising from defining equations. As a result, trapezoidal algorithm has suboptimal parallelism.

State-of-the-art cache-oblivious parallel algorithms for dynamic programming (DP) problems usually guarantee asymptotically optimal cache performance without any tuning of parameters, but they often fail to exploit the theoretically best parallelism at same time. While these achieve cache-optimality through use a recursive divide-and-conquer (DAC) strategy, scheduling tasks granularity task dependency introduces artificial dependencies in addition those arising from defining recurrence...

Given a polygonal region containing target point (which we assume is the origin), it not hard to see that there are two points on perimeter antipodal, i.e., whose midpoint origin. We prove three generalizations of this fact. (1) For any polygon (or bounded closed with connected boundary) origin, possible place given set weights boundary so their barycenter (center mass) coincides provided largest weight does exceed sum other weights. (2) On 3-dimensional polyhedron exist form an equilateral...

We present AUTOGEN---an algorithm that for a wide class of dynamic programming (DP) problems automatically discovers highly efficient cache-oblivious parallel recursive divide-and-conquer algorithms from inefficient iterative descriptions DP recurrences. AUTOGEN analyzes the set table locations accessed by when run on small size, and identifies access pattern corresponding provably correct solving recurrence. use to autodiscover several well-known problems. Our experimental results show...

Smoke fog or other light-interference environments have intrinsic obstruction for visual recognition techniques to explore objects and surroundings. Alternatively, tactile perceptions, rather than observations, are commonly used by burrowing deep-sea animals communicate with environments. Bio-inspired this natural wisdom, here, we demonstrate stretchable tentacle sensor arrays, which can recognize surrounding located in non-visual conditions such as smoke dark environment. Each is composed...

Existing point cloud semantic segmentation networks cannot identify unknown classes and update their knowledge, due to a closed-set static perspective of the real world, which would induce intelligent agent make bad decisions. To address this problem, we propose Probability-Driven Framework (PDF) for open world that includes (i) lightweight U-decoder branch by estimating uncertainties, (ii) flexible pseudo-labeling scheme supply geometry features along with probability distribution...

<title>Abstract</title> A microgrid with small hydropower as the main power source can absorb or transmit to grid during grid-connected operation. When is operating off-grid due sudden reasons such line faults, it necessary remove sources loads maintain frequency stability of microgrid. Due difficulty achieving high-quality communication in mountainous areas where concentrated, a common solution use high-frequency generation tripping low-frequency load shedding without communication. This...

Large 2D vision-language models (2D-LLMs) have gained significant attention by bridging Language Models (LLMs) with images using a simple projector. Inspired their success, large 3D point cloud-language (3D-LLMs) also integrate clouds into LLMs. However, directly aligning LLM requires expensive training costs, typically in hundreds of GPU-hours on A100, which hinders the development 3D-LLMs. In this paper, we introduce MiniGPT-3D, an efficient and powerful 3D-LLM that achieves multiple SOTA...

Point scene instance mesh reconstruction is a challenging task since it requires both scene-level segmentation and instance-level from partial observations simultaneously. Previous works either adopt detection backbone or one, then directly employ network to produce complete meshes incomplete point clouds. To further boost the quality with local details global smoothness, in this work, we propose JIMR, joint framework two cascaded stages for semantic geometry understanding. In first stage,...