Semantic IDs (SIDs) are hierarchical item identifiers learned via residual quantization, offering a promising solution to cold-start recommendation, yet their integration into discriminative recommendations remains challenging and lacks systematic investigation. We present Pin-SCALE, a framework for optimized end-to-end integration of SID into dense retrieval in cascading recommender systems. Pin-SCALE instantiates the principles through three technical contributions: (1) cascading pooling preserving residual quantization hierarchy that outperforms common pooling schema and attention;(2) engagement-aware tokenization that encodes collaborative signals to fully utilize both content and user engagement pattern; and (3) multi-view contrastive learning that harnesses query-item and item-item co-occurrence to align SID representations with engagement prediction. Through extensive offline analyses, we collected systematic guidance and proved the effectiveness of Pin-SCALE. Furthermore, the framework has been tested online on multiple surfaces at Pinterest including Closeup, Home Feed and Search, with different content types like organic and e-commerce. Pin-SCALE achieves substantial improvements in fresh engagements (+3.67% repins) as well as platform-level retention (+0.05% DAU). Our work has been shipped into production to serve billions of requests daily and drive discovery journey of our users, as well as provides principled guidelines for SID adoption in discriminative recommendation models at production scale.

Pinterest is the visual discovery platform where people find inspiration, curate ideas, and shop products for all life's moments. An intentful journey can start when Pinners (users) click on Pins and arrive at the Closeup surface, where they can continue to explore or refine their intent by browsing related content powered by our visual search and related recommendation engines. Product Pins, or contents that are linked to merchants and are buyable in general, are critical to realizable fulfillment in Pinners' exploratory journey. This paper focuses on optimizing embedding-based retrieval (EBR) to retrieve relevant and personalized product Pins to drive actionable engagement. In contrast to conventional EBR systems, we introduce complementary Shopping Priority Corpora that are prepared by probability models from a dynamic inventory with billions of candidates and highly skewed distributions. This novel design enabled us to significantly improve the retrieval efficiency, scale up the systems, and meet different business requirements. On top of that, we build retrieval models that infuse multimodal information with multi-task contrastive learning to balance relevance and engagement. We evaluate the EBR system on random off-policy traffic with thorough baseline comparisons and rigorous online A/B experiments. This work leads to significant metric gains in our production systems and provides practical lessons on improving early retrieval for multiple business objectives at large scales.

In recent years, with the increasing demand for computing power from various intelligent applications on mobile devices, heterogeneous multi-core architectures have received more attention. The big.LITTLE architecture, as a new type of energy-saving architecture, consists of high-performance cores and low energy consumption cores, which can better meet the real-time task requirements of various mobile devices. One of the challenges faced by these scheduling algorithms is how to fully utilize the advantages of big.LITTLE architectures to achieve a balance between system energy consumption and performance. To address this issue, we designed an initial task allocation strategy based on the characteristics of big.LITTLE architecture, introduced reinforcement learning algorithms for decision-making, and assigned different types of tasks to their appropriate processors. In order to balance task load and reduce total scheduling time, we proposed a dynamic migration algorithm, which dynamically adjusts task load by migrating them in the case of imbalanced utilization. Experimental results demonstrate that the proposed algorithm effectively balances energy consumption and performance, improves resource utilization, and reduces both scheduling time and energy consumption.

Task offloading is a key technology in fog computing, which allows resource-intensive tasks to be offloaded from terminal devices to fog nodes with higher computing capabilities. However, traditional task offloading methods usually use heuristic methods, which are highly dependent on the mode. They cannot effectively optimize the delay, resulting in a decline in service quality. To solve this problem, a task offloading method based on deep reinforcement learning (DRL) is proposed in this paper. We applied various methods to improve the Deep Q-Network (DQN) and utilized it for task offloading to enhance system performance in fog computing.

Software architectures are considered the key means to manage the complexity of large-scale systems from the high abstraction levels and system-wide perspectives. The traditional software design methodologies and the emerging architecture design methods still fall short of coping with the architectural complexity and difficulty in practice. The recent research on the architecture design decisions mostly focuses on its representation, providing little support for the architecture design task itself. In this paper we propose a decision-oriented architecture design approach ABC/DD, based on the decision-abstraction and issue-decomposition principles specific to the architecture level design of software. The approach models software architecture from the perspective of design decisions, and accomplishes the architecture design from eliciting architecturally significant design issues to exploiting and making decisions on the solutions for these issues. We illustrate the application of the approach with two real-life large-scale software-intensive projects, showing that the decision-oriented approach accommodates the characteristics and demands of the architecture level, and facilitates the design of architecture and the capture of the essential decisions for large complex systems.