Mark Zhao

Mark Zhao, Dhruv Choudhary, Devashish Tyagi et al. · stanford

We present RecD (Recommendation Deduplication), a suite of end-to-end infrastructure optimizations across the Deep Learning Recommendation Model (DLRM) training pipeline. RecD addresses immense storage, preprocessing, and training overheads caused by feature duplication inherent in industry-scale DLRM training datasets. Feature duplication arises because DLRM datasets are generated from interactions. While each user session can generate multiple training samples, many features' values do not change across these samples. We demonstrate how RecD exploits this property, end-to-end, across a deployed training pipeline. RecD optimizes data generation pipelines to decrease dataset storage and preprocessing resource demands and to maximize duplication within a training batch. RecD introduces a new tensor format, InverseKeyedJaggedTensors (IKJTs), to deduplicate feature values in each batch. We show how DLRM model architectures can leverage IKJTs to drastically increase training throughput. RecD improves the training and preprocessing throughput and storage efficiency by up to 2.48x, 1.79x, and 3.71x, respectively, in an industry-scale DLRM training system.

Rui Yang, Hanyang Chen, Junyu Zhang et al.

Leveraging Multi-modal Large Language Models (MLLMs) to create embodied agents offers a promising avenue for tackling real-world tasks. While language-centric embodied agents have garnered substantial attention, MLLM-based embodied agents remain underexplored due to the lack of comprehensive evaluation frameworks. To bridge this gap, we introduce EmbodiedBench, an extensive benchmark designed to evaluate vision-driven embodied agents. EmbodiedBench features: (1) a diverse set of 1,128 testing tasks across four environments, ranging from high-level semantic tasks (e.g., household) to low-level tasks involving atomic actions (e.g., navigation and manipulation); and (2) six meticulously curated subsets evaluating essential agent capabilities like commonsense reasoning, complex instruction understanding, spatial awareness, visual perception, and long-term planning. Through extensive experiments, we evaluated 24 leading proprietary and open-source MLLMs within EmbodiedBench. Our findings reveal that: MLLMs excel at high-level tasks but struggle with low-level manipulation, with the best model, GPT-4o, scoring only 28.9\% on average. EmbodiedBench provides a multifaceted standardized evaluation platform that not only highlights existing challenges but also offers valuable insights to advance MLLM-based embodied agents. Our code and dataset are available at https://embodiedbench.github.io.

Junyoung Kim, Anton Korikov, Jiazhou Liang et al.

While Large Language Models (LLMs) exhibit exceptional zero-shot relevance modeling, their high computational cost necessitates framing passage retrieval as a budget-constrained global optimization problem. Existing approaches passively rely on first-stage dense retrievers, which leads to two limitations: (1) failing to retrieve relevant passages in semantically distinct clusters, and (2) failing to propagate relevance signals to the broader corpus. To address these limitations, we propose Bayesian Active Learning with Gaussian Processes guided by LLM relevance scoring (BAGEL), a novel framework that propagates sparse LLM relevance signals across the embedding space to guide global exploration. BAGEL models the multimodal relevance distribution across the entire embedding space with a query-specific Gaussian Process (GP) based on LLM relevance scores. Subsequently, it iteratively selects passages for scoring by strategically balancing the exploitation of high-confidence regions with the exploration of uncertain areas. Extensive experiments across four benchmark datasets and two LLM backbones demonstrate that BAGEL effectively explores and captures complex relevance distributions and outperforms LLM reranking methods under the same LLM budget on all four datasets.

Swapnil Gandhi, Mark Zhao, Athinagoras Skiadopoulos et al.

Training large Deep Neural Network (DNN) models requires thousands of GPUs over the course of several days or weeks. At this scale, failures are frequent and can have a big impact on training throughput. Utilizing spare GPU servers to mitigate performance loss becomes increasingly costly as model sizes grow. ReCycle is a system designed for efficient DNN training in the presence of failures, without relying on spare servers. It exploits the inherent functional redundancy in distributed training systems -- where servers across data-parallel groups store the same model parameters -- and pipeline schedule bubbles within each data-parallel group. When servers fails, ReCycle dynamically re-routes micro-batches to data-parallel peers, allowing for uninterrupted training despite multiple failures. However, this re-routing can create imbalances across pipeline stages, leading to reduced training throughput. To address this, ReCycle introduces two key optimizations that ensure re-routed micro-batches are processed within the original pipeline schedule's bubbles. First, it decouples the backward pass into two phases: one for computing gradients for the input and another for calculating gradients for the parameters. Second, it avoids synchronization across pipeline stages by staggering the optimizer step. Together, these optimizations enable adaptive pipeline schedules that minimize or even eliminate training throughput degradation during failures. We describe a prototype for ReCycle and show that it achieves high training throughput under multiple failures, outperforming recent proposals for fault-tolerant training such as Oobleck and Bamboo by up to $1.46\times$ and $1.64\times$, respectively.

Mark Zhao, Emanuel Adamiak, Christos Kozyrakis

The input data pipeline is an essential component of each machine learning (ML) training job. It is responsible for reading massive amounts of training data, processing batches of samples using complex transformations, and loading them onto training nodes at low latency and high throughput. Performant input data systems are becoming increasingly critical, driven by skyrocketing data volumes and training throughput demands. Unfortunately, current input data systems cannot fully leverage key performance optimizations, resulting in hugely inefficient infrastructures that require significant resources - or worse - underutilize expensive accelerators. To address these demands, we present cedar, an optimized and unified programming framework for ML input data pipelines. cedar allows users to define input data pipelines using composable operators that support arbitrary ML frameworks and libraries. cedar introduces an extensible optimizer that systematically applies a complex combination of optimizations (e.g., offloading, caching, prefetching, fusion, and reordering). It orchestrates processing across a customizable set of local and distributed compute resources in order to improve processing performance and efficiency, all without user input. Across eight pipelines, cedar improves performance by up to 1.87x to 10.65x compared to state-of-the-art input data systems.

Hanyang Chen, Mark Zhao, Rui Yang et al.

Recent advances in embodied AI highlight the potential of vision language models (VLMs) as agents capable of perception, reasoning, and interaction in complex environments. However, top-performing systems rely on large-scale models that are costly to deploy, while smaller VLMs lack the necessary knowledge and skills to succeed. To bridge this gap, we present \textit{Embodied Reasoning Agent (ERA)}, a two-stage framework that integrates prior knowledge learning and online reinforcement learning (RL). The first stage, \textit{Embodied Prior Learning}, distills foundational knowledge from three types of data: (1) Trajectory-Augmented Priors, which enrich existing trajectory data with structured reasoning generated by stronger models; (2) Environment-Anchored Priors, which provide in-environment knowledge and grounding supervision; and (3) External Knowledge Priors, which transfer general knowledge from out-of-environment datasets. In the second stage, we develop an online RL pipeline that builds on these priors to further enhance agent performance. To overcome the inherent challenges in agent RL, including long horizons, sparse rewards, and training instability, we introduce three key designs: self-summarization for context management, dense reward shaping, and turn-level policy optimization. Extensive experiments on both high-level planning (EB-ALFRED) and low-level control (EB-Manipulation) tasks demonstrate that ERA-3B surpasses both prompting-based large models and previous training-based baselines. Specifically, it achieves overall improvements of 8.4\% on EB-ALFRED and 19.4\% on EB-Manipulation over GPT-4o, and exhibits strong generalization to unseen tasks. Overall, ERA offers a practical path toward scalable embodied intelligence, providing methodological insights for future embodied AI systems.

Athinagoras Skiadopoulos, Mark Zhao, Swapnil Gandhi et al.

Mixture-of-Experts (MoE) models have become a widely-adopted solution to continue scaling model sizes without a corresponding linear increase in compute. During MoE model training, each input token is dynamically routed to a subset of experts -- sparsely-activated feed-forward networks -- within each transformer layer. The distribution of tokens assigned to each expert varies widely and rapidly over the course of training. To handle the wide load imbalance across experts, current systems are forced to either drop tokens assigned to popular experts, degrading convergence, or frequently rebalance resources allocated to each expert based on popularity, incurring high state migration overheads. To break this performance-accuracy tradeoff, we introduce SYMI, an adaptive MoE training system. The key insight of SYMI is to decouple the placement of expert parameters from their large optimizer state. SYMI statically partitions the optimizer of each expert across all training nodes. Meanwhile, SYMI dynamically adjusts the placement of expert parameters by repurposing existing weight updates, avoiding migration overheads. In doing so, SYMI right-sizes the GPU resources allocated to each expert, on a per-iteration basis, with minimal overhead. Compared to state-of-the-art MoE training systems, DeepSpeed and FlexMoE, SYMI is able to achieve a 30.5% and 25.9% faster time-to-convergence, respectively.

Luis Gaspar Schroeder, Aditya Desai, Alejandro Cuadron et al.

Semantic caches return cached responses for semantically similar prompts to reduce LLM inference latency and cost. They embed cached prompts and store them alongside their response in a vector database. Embedding similarity metrics assign a numerical score to quantify the similarity between a request and its nearest neighbor prompt from the cache. Existing systems use the same static similarity threshold across all requests to determine whether two prompts can share similar responses. However, we observe that static thresholds do not give formal correctness guarantees, can result in unexpected error rates, and lead to suboptimal cache hit rates. This paper proposes vCache, the first verified semantic cache with user-defined error rate guarantees. It employs an online learning algorithm to estimate an optimal threshold for each cached prompt, enabling reliable cache responses without additional training. Our experiments show that vCache consistently meets the specified error bounds while outperforming state-of-the-art static-threshold and fine-tuned embedding baselines. We release the vCache implementation and three benchmarks to support future research.

Mark Zhao, Niket Agarwal, Aarti Basant et al.

Datacenter-scale AI training clusters consisting of thousands of domain-specific accelerators (DSA) are used to train increasingly-complex deep learning models. These clusters rely on a data storage and ingestion (DSI) pipeline, responsible for storing exabytes of training data and serving it at tens of terabytes per second. As DSAs continue to push training efficiency and throughput, the DSI pipeline is becoming the dominating factor that constrains the overall training performance and capacity. Innovations that improve the efficiency and performance of DSI systems and hardware are urgent, demanding a deep understanding of DSI characteristics and infrastructure at scale. This paper presents Meta's end-to-end DSI pipeline, composed of a central data warehouse built on distributed storage and a Data PreProcessing Service that scales to eliminate data stalls. We characterize how hundreds of models are collaboratively trained across geo-distributed datacenters via diverse and continuous training jobs. These training jobs read and heavily filter massive and evolving datasets, resulting in popular features and samples used across training jobs. We measure the intense network, memory, and compute resources required by each training job to preprocess samples during training. Finally, we synthesize key takeaways based on our production infrastructure characterization. These include identifying hardware bottlenecks, discussing opportunities for heterogeneous DSI hardware, motivating research in datacenter scheduling and benchmark datasets, and assimilating lessons learned in optimizing DSI infrastructure.

Mark Zhao, Mingyu Gao, Christos Kozyrakis

FPGAs are now used in public clouds to accelerate a wide range of applications, including many that operate on sensitive data such as financial and medical records. We present ShEF, a trusted execution environment (TEE) for cloud-based reconfigurable accelerators. ShEF is independent from CPU-based TEEs and allows secure execution under a threat model where the adversary can control all software running on the CPU connected to the FPGA, has physical access to the FPGA, and can compromise the FPGA interface logic of the cloud provider. ShEF provides a secure boot and remote attestation process that relies solely on existing FPGA mechanisms for root of trust. It also includes a Shield component that provides secure access to data while the accelerator is in use. The Shield is highly customizable and extensible, allowing users to craft a bespoke security solution that fits their accelerator's memory access patterns, bandwidth, and security requirements at minimum performance and area overheads. We describe a prototype implementation of ShEF for existing cloud FPGAs, map ShEF to a performant and secure storage application, and measure the performance benefits of customizable security using five additional accelerators.