MAROKO133 Breaking ai: Taiwan commits $250B to US semiconductor chip plants in exchange fo

📌 MAROKO133 Eksklusif ai: Taiwan commits $250B to US semiconductor chip plants in

The United States and Taiwan have reached a major trade agreement aimed at expanding semiconductor manufacturing on American soil.

The U.S. Department of Commerce announced the deal Thursday, outlining large-scale investment commitments and tariff adjustments tied to chip production in the United States.

Under the agreement, Taiwanese chip and technology firms will invest at least $250 billion in U.S. production capacity.

The Taiwanese government will also guarantee $250 billion in credit to support these companies.

In return, the United States will cap “reciprocal” tariffs on Taiwan at 15%, down from 20 percent.

The deal also sets zero reciprocal tariffs on generic pharmaceuticals, their ingredients, aircraft components, and selected natural resources.

Commerce Secretary Howard Lutnick said the agreement could accelerate new chip projects in Arizona.

He confirmed that Taiwan Semiconductor Manufacturing Co. has already secured additional land near its existing site.

“They just bought hundreds of acres adjacent to their property,” Lutnick said in an interview with CNBC. “I’ll let them go through with their board and give them time.”

Investment tied to tariffs

The agreement links tariff relief directly to U.S. manufacturing activity.

The Commerce Department said future tariffs under the Section 232 framework will include exceptions for firms building chips in the United States.

While new fabs are under construction, Taiwanese companies will be able to import up to 2.5 times the capacity they are building without paying Section 232 tariffs.

Once factories begin operations, companies can import up to 1.5 times their U.S. production capacity.

Taiwanese auto parts, lumber, and related products will also avoid tariffs above 15 percent under the same framework.

Lutnick warned that companies refusing to invest in U.S. manufacturing could face severe penalties.

“That’s what they get if they don’t build in America, the tariff’s likely to be 100 percent,” he said.

He added that Washington aims to relocate 40 percent of Taiwan’s semiconductor supply chain to the United States.

TSMC expansion outlook

TSMC, the world’s largest contract chipmaker, said it continues to weigh its investment plans based on market demand.

“Regarding TSMC’s plans, the market demand for our advanced technology is very strong, we continue to invest in Taiwan and expand overseas, all the investment decisions are based on market conditions and customer demands,” a TSMC spokesperson told CNBC.

The company has already committed as much as $40 billion to build chip fabs in Arizona.

Those facilities will produce chips for major U.S. clients, including Apple and NVIDIA, using funding support from the CHIPS Act.

TSMC welcomed the broader trade framework between Washington and Taipei.

“As a semiconductor foundry serving customers worldwide, we welcome the prospect of robust trade agreements between the United States and Taiwan,” a company representative said.

U.S. officials have framed the agreement as a matter of economic and national security.

The government has prioritized domestic production of leading-edge chips as competition over artificial intelligence hardware intensifies.

Officials have also warned that any disruption to Taiwan’s chip industry, including a potential conflict with China, could damage the U.S. economy.

“We’re going to bring it all over so we become self-sufficient in the capacity of building semiconductors,” Lutnick said.

🔗 Sumber: interestingengineering.com

📌 MAROKO133 Eksklusif ai: ByteDance Introduces Astra: A Dual-Model Architecture fo

The increasing integration of robots across various sectors, from industrial manufacturing to daily life, highlights a growing need for advanced navigation systems. However, contemporary robot navigation systems face significant challenges in diverse and complex indoor environments, exposing the limitations of traditional approaches. Addressing the fundamental questions of “Where am I?”, “Where am I going?”, and “How do I get there?”, ByteDance has developed Astra, an innovative dual-model architecture designed to overcome these traditional navigation bottlenecks and enable general-purpose mobile robots.

Traditional navigation systems typically consist of multiple, smaller, and often rule-based modules to handle the core challenges of target localization, self-localization, and path planning. Target localization involves understanding natural language or image cues to pinpoint a destination on a map. Self-localization requires a robot to determine its precise position within a map, especially challenging in repetitive environments like warehouses where traditional methods often rely on artificial landmarks (e.g., QR codes). Path planning further divides into global planning for rough route generation and local planning for real-time obstacle avoidance and reaching intermediate waypoints.

While foundation models have shown promise in integrating smaller models to tackle broader tasks, the optimal number of models and their effective integration for comprehensive navigation remained an open question.

ByteDance’s Astra, detailed in their paper “Astra: Toward General-Purpose Mobile Robots via Hierarchical Multimodal Learning” (website: https://astra-mobility.github.io/), addresses these limitations. Following the System 1/System 2 paradigm, Astra features two primary sub-models: Astra-Global and Astra-Local. Astra-Global handles low-frequency tasks like target and self-localization, while Astra-Local manages high-frequency tasks such as local path planning and odometry estimation. This architecture promises to revolutionize how robots navigate complex indoor spaces.

Astra-Global: The Intelligent Brain for Global Localization

Astra-Global serves as the intelligent core of the Astra architecture, responsible for critical low-frequency tasks: self-localization and target localization. It functions as a Multimodal Large Language Model (MLLM), adept at processing both visual and linguistic inputs to achieve precise global positioning within a map. Its strength lies in utilizing a hybrid topological-semantic graph as contextual input, allowing the model to accurately locate positions based on query images or text prompts.

The construction of this robust localization system begins with offline mapping. The research team developed an offline method to build a hybrid topological-semantic graph G=(V,E,L):

• V (Nodes): Keyframes, obtained by temporal downsampling of input video and SfM-estimated 6-Degrees-of-Freedom (DoF) camera poses, act as nodes encoding camera poses and landmark references.
• E (Edges): Undirected edges establish connectivity based on relative node poses, crucial for global path planning.
• L (Landmarks): Semantic landmark information is extracted by Astra-Global from visual data at each node, enriching the map’s semantic understanding. These landmarks store semantic attributes and are connected to multiple nodes via co-visibility relationships.

In practical localization, Astra-Global’s self-localization and target localization capabilities leverage a coarse-to-fine two-stage process for visual-language localization. The coarse stage analyzes input images and localization prompts, detects landmarks, establishes correspondence with a pre-built landmark map, and filters candidates based on visual consistency. The fine stage then uses the query image and coarse output to sample reference map nodes from the offline map, comparing their visual and positional information to directly output the predicted pose.

For language-based target localization, the model interprets natural language instructions, identifies relevant landmarks using their functional descriptions within the map, and then leverages landmark-to-node association mechanisms to locate relevant nodes, retrieving target images and 6-DoF poses.

To empower Astra-Global with robust localization abilities, the team employed a meticulous training methodology. Using Qwen2.5-VL as the backbone, they combined Supervised Fine-Tuning (SFT) with Group Relative Policy Optimization (GRPO). SFT involved diverse datasets for various tasks, including coarse and fine localization, co-visibility detection, and motion trend estimation. In the GRPO phase, a rule-based reward function (including format, landmark extraction, map matching, and extra landmark rewards) was used to train for visual-language localization. Experiments showed GRPO significantly improved Astra-Global’s zero-shot generalization, achieving 99.9% localization accuracy in unseen home environments, surpassing SFT-only methods.

Astra-Local: The Intelligent Assistant for Local Planning

Astra-Local acts as the intelligent assistant for Astra’s high-frequency tasks, a multi-task network capable of efficiently generating local paths and accurately estimating odometry from sensor data. Its architecture comprises three core components: a 4D spatio-temporal encoder, a planning head, and an odometry head.

The 4D spatio-temporal encoder replaces traditional mobile stack perception and prediction modules. It begins with a 3D spatial encoder that processes N omnidirectional images through a Vision Transformer (ViT) and Lift-Splat-Shoot to convert 2D image features into 3D voxel features. This 3D encoder is trained using self-supervised learning via 3D volumetric differentiable neural rendering. The 4D spatio-temporal encoder then builds upon the 3D encoder, taking past voxel features and future timestamps as input to predict future voxel features through ResNet and DiT modules, providing current and future environmental representations for planning and odometry.

The planning head, based on pre-trained 4D features, robot speed, and task information, generates executable trajectories using Transformer-based flow matching. To prevent collisions, the planning head incorporates a masked ESDF loss (Euclidean Signed Distance Field). This loss calculates the ESDF of a 3D occupancy map and applies a 2D ground truth trajectory mask, significantly reducing collision rates. Experiments demonstrate its superior performance in collision rate and overall score on out-of-distribution (OOD) datasets compared to other methods.

The odometry head predicts the robot’s relative pose using current and past 4D features and additional sensor data (e.g., IMU, wheel data). It trains a Transformer model to fuse information from different sensors. Each sensor modality is processed by a specific tokenizer, combined with modality embeddings and temporal positional embeddi…

Konten dipersingkat otomatis.

🔗 Sumber: syncedreview.com