arXiv subjects

Pingzhi Li

Publications and source records attributed to Pingzhi Li.

At least 37 records · Page 2Linked to original sources

arXiv2025

Occult: Optimizing Collaborative Communication across Experts for Accelerated Parallel MoE Training and Inference

Mixture-of-experts (MoE) architectures could achieve impressive computational efficiency with expert parallelism, which relies heavily on all-to-all communication across devices. Unfortunately, such communication overhead typically constitutes a significant portion of the total runtime, hampering the scalability of distributed training and inference for modern MoE models (consuming over $40\%$ runtime in large-scale training). In this paper, we first define collaborative communication to illustrate this intrinsic limitation, and then propose system- and algorithm-level innovations to reduce communication costs. Specifically, given a pair of experts co-activated by one token, we call them "collaborated", which comprises $2$ cases as intra- and inter-collaboration, depending on whether they are kept on the same device. Our pilot investigations reveal that augmenting the proportion of intra-collaboration can accelerate expert parallelism at scale. It motivates us to strategically optimize collaborative communication for accelerated MoE training and inference, dubbed Occult. Our designs are capable of either delivering exact results with reduced communication cost or controllably minimizing the cost with collaboration pruning, materialized by modified fine-tuning. Comprehensive experiments on various MoE-LLMs demonstrate that Occult can be faster than popular state-of-the-art inference or training frameworks (more than $1.5\times$ speed up across multiple tasks and models) with comparable or superior quality compared to the standard fine-tuning. Code is available at $\href{https://github.com/UNITES-Lab/Occult}{https://github.com/UNITES-Lab/Occult}$.

cs.LG↗

arXiv2025

Q-Newton: Hybrid Quantum-Classical Scheduling for Accelerating Neural Network Training with Newton's Gradient Descent

Optimization techniques in deep learning are predominantly led by first-order gradient methodologies, such as SGD. However, neural network training can greatly benefit from the rapid convergence characteristics of second-order optimization. Newton's GD stands out in this category, by rescaling the gradient using the inverse Hessian. Nevertheless, one of its major bottlenecks is matrix inversion, which is notably time-consuming in $O(N^3)$ time with weak scalability. Matrix inversion can be translated into solving a series of linear equations. Given that quantum linear solver algorithms (QLSAs), leveraging the principles of quantum superposition and entanglement, can operate within a $\text{polylog}(N)$ time frame, they present a promising approach with exponential acceleration. Specifically, one of the most recent QLSAs demonstrates a complexity scaling of $O(d\cdotκ\log(N\cdotκ/ε))$, depending on: {size~$N$, condition number~$κ$, error tolerance~$ε$, quantum oracle sparsity~$d$} of the matrix. However, this also implies that their potential exponential advantage may be hindered by certain properties (i.e. $κ$ and $d$). We propose Q-Newton, a hybrid quantum-classical scheduler for accelerating neural network training with Newton's GD. Q-Newton utilizes a streamlined scheduling module that coordinates between quantum and classical linear solvers, by estimating & reducing $κ$ and constructing $d$ for the quantum solver. Our evaluation showcases the potential for Q-Newton to significantly reduce the total training time compared to commonly used optimizers like SGD. We hypothesize a future scenario where the gate time of quantum machines is reduced, possibly realized by attoseconds physics. Our evaluation establishes an ambitious and promising target for the evolution of quantum computing.

quant-ph↗

arXiv2025

Advancing MoE Efficiency: A Collaboration-Constrained Routing (C2R) Strategy for Better Expert Parallelism Design

Mixture-of-Experts (MoE) has successfully scaled up models while maintaining nearly constant computing costs. By employing a gating network to route input tokens, it selectively activates a subset of expert networks to process the corresponding token embeddings. However, in practice, the efficiency of MoE is challenging to achieve due to two key reasons: imbalanced expert activation, which leads to substantial idle time during model or expert parallelism, and insufficient capacity utilization; massive communication overhead, induced by numerous expert routing combinations in expert parallelism at the system level. Previous works typically formulate it as the load imbalance issue characterized by the gating network favoring certain experts over others or attribute it to static execution which fails to adapt to the dynamic expert workload at runtime. In this paper, we exploit it from a brand new perspective, a higher-order view and analysis of MoE routing policies: expert collaboration and specialization where some experts tend to activate broadly with others (collaborative), while others are more likely to activate only with a specific subset of experts (specialized). Our experiments reveal that most experts tend to be overly collaborative, leading to increased communication overhead from repeatedly sending tokens to different accelerators. To this end, we propose a novel collaboration-constrained routing (C2R) strategy to encourage more specialized expert groups, as well as to improve expert utilization, and present an efficient implementation of MoE that further leverages expert specialization. We achieve an average performance improvement of 0.51% and 0.33% on LLaMA-MoE and Qwen-MoE respectively across ten downstream NLP benchmarks, and reduce the all2all communication costs between GPUs, bringing an extra 20%-30% total running time savings on top of the existing SoTA, i.e. MegaBlocks.

cs.LG↗

arXiv2025

Glider: Global and Local Instruction-Driven Expert Router

The availability of performant pre-trained models has led to a proliferation of fine-tuned expert models that are specialized to particular domains. This has enabled the creation of powerful and adaptive routing-based "Model MoErging" methods with the goal of using expert modules to create an aggregate system with improved performance or generalization. However, existing MoErging methods often prioritize generalization to unseen tasks at the expense of performance on held-in tasks, which limits its practical applicability in real-world deployment scenarios. We observe that current token-level routing mechanisms neglect the global semantic context of the input task. This token-wise independence hinders effective expert selection for held-in tasks, as routing decisions fail to incorporate the semantic properties of the task. To address this, we propose, Global and Local Instruction Driven Expert Router (GLIDER) that integrates a multi-scale routing mechanism, encompassing a semantic global router and a learned local router. The global router leverages LLM's advanced reasoning capabilities for semantic-related contexts to enhance expert selection. Given the input query and LLM, the router generates semantic task instructions that guide the retrieval of the most relevant experts across all layers. This global guidance is complemented by a local router that facilitates token-level routing decisions within each module, enabling finer control and enhanced performance on unseen tasks. Our experiments using T5-based models for T0 and FLAN tasks demonstrate that GLIDER achieves substantially improved held-in performance while maintaining strong generalization on held-out tasks. We also perform ablations experiments to dive deeper into the components of GLIDER. Our experiments highlight the importance of our multi-scale routing that leverages LLM-driven semantic reasoning for MoErging methods.

cs.LG↗

arXiv2025

Finding Fantastic Experts in MoEs: A Unified Study for Expert Dropping Strategies and Observations

Sparsely activated Mixture-of-Experts (SMoE) has shown promise in scaling up the learning capacity of neural networks. However, vanilla SMoEs have issues such as expert redundancy and heavy memory requirements, making them inefficient and non-scalable, especially for resource-constrained scenarios. Expert-level sparsification of SMoEs involves pruning the least important experts to address these limitations. In this work, we aim to address three questions: (1) What is the best recipe to identify the least knowledgeable subset of experts that can be dropped with minimal impact on performance? (2) How should we perform expert dropping (one-shot or iterative), and what correction measures can we undertake to minimize its drastic impact on SMoE subnetwork capabilities? (3) What capabilities of full-SMoEs are severely impacted by the removal of the least dominant experts, and how can we recover them? Firstly, we propose MoE Experts Compression Suite (MC-Suite), which is a collection of some previously explored and multiple novel recipes to provide a comprehensive benchmark for estimating expert importance from diverse perspectives, as well as unveil numerous valuable insights for SMoE experts. Secondly, unlike prior works with a one-shot expert pruning approach, we explore the benefits of iterative pruning with the re-estimation of the MC-Suite criterion. Moreover, we introduce the benefits of task-agnostic fine-tuning as a correction mechanism during iterative expert dropping, which we term MoE Lottery Subnetworks. Lastly, we present an experimentally validated conjecture that, during expert dropping, SMoEs' instruction-following capabilities are predominantly hurt, which can be restored to a robust level subject to external augmentation of instruction-following capabilities using k-shot examples and supervised fine-tuning.

cs.LG↗

arXiv2025

ORAL: Prompting Your Large-Scale LoRAs via Conditional Recurrent Diffusion

Parameter generation has emerged as a novel paradigm for neural network development, offering an alternative to traditional neural network training by synthesizing high-quality model weights directly. In the context of Low-Rank Adaptation (LoRA) for evolving ($\textit{i.e.}$, constantly updated) large language models (LLMs), this approach promises efficient adaptation without costly retraining. However, existing methods face critical limitations in simultaneously achieving scalability and controllability. In this paper, we introduce $\texttt{ORAL}$, a novel $\textbf{conditional recurrent diffusion}$ framework that addresses these challenges. $\texttt{ORAL}$ incorporates a novel conditioning mechanism that integrates model architecture and textual task specifications, enabling the generation of task-specific LoRA parameters that can seamlessly transfer across evolving foundation models. Our approach successfully scales to billions-of-parameter LLMs and maintains controllability. Through extensive experiments across seven language tasks, four vision tasks, and three multimodal tasks using five pre-trained LLMs, we demonstrate that $\texttt{ORAL}$ generates high-quality LoRA parameters that achieve comparable or superior performance to vanilla trained counterparts.

cs.LG↗

arXiv2025

Hexa-MoE: Efficient and Heterogeneous-aware Training for Mixture-of-Experts

Mixture-of-Experts (MoE) has emerged as a practical approach to scale up parameters for the Transformer model to achieve better generalization while maintaining a sub-linear increase in computation overhead. Current MoE models are mainly built with expert parallelism on distributed devices. However, it usually depends on homogeneous devices to deploy and suffers from heavy communication overhead and computation redundancy. In this paper, we explore developing a \texttt{H}eterogeneous-aware \texttt{EX}pert \texttt{A}llocation framework, \textbf{\texttt{HEXA-MoE}}, with significantly enhanced computing efficiency. It contains two components: ($1$) \textit{Expert-Specific Operators}. We replace the typical general matrix multiplication or grouped matrix multiplication interfaces with our operators, which allows the computing to be performed in an in-place manner with \textbf{ZERO} redundancy. ($2$) \textit{Adaptive Data- and Model-Centric Configurations} for different workload scales. Specifically, we introduce a pipeline-shared cache on each device to tackle the heavy memory consumption in the existing data-centric MoE library. Comprehensive experiments on the Swin-MoE benchmark consistently reveal the effectiveness of our \texttt{HEXA-MoE} framework, i.e., reducing $10\%\sim48\%$ memory consumption and achieving $0.5\sim4.3\times$ speed up compared to current state-of-the-art MoE libraries. Furthermore, we examine our \texttt{HEXA-MoE} with heterogeneous devices for both data- and model-centric settings. Promising results show that employing optimal parallel configuration with \texttt{HEXA-MoE} on heterogeneous devices can substantially minimize overall latency. Codes are available at https://github.com/UNITES-Lab/HEXA-MoE.

cs.DC↗

arXiv2025

PortLLM: Personalizing Evolving Large Language Models with Training-Free and Portable Model Patches

As large language models (LLMs) increasingly shape the AI landscape, fine-tuning pretrained models has become more popular than in the pre-LLM era for achieving optimal performance in domain-specific tasks. However, pretrained LLMs such as ChatGPT are periodically evolved, i.e., model parameters are frequently updated), making it challenging for downstream users with limited resources to keep up with fine-tuning the newest LLMs for their domain application. Even though fine-tuning costs have nowadays been reduced thanks to the innovations of parameter-efficient fine-tuning such as LoRA, not all downstream users have adequate computing for frequent personalization. Moreover, access to fine-tuning datasets, particularly in sensitive domains such as healthcare, could be time-restrictive, making it crucial to retain the knowledge encoded in earlier fine-tuned rounds for future adaptation. In this paper, we present PortLLM, a training-free framework that (i) creates an initial lightweight model update patch to capture domain-specific knowledge, and (ii) allows a subsequent seamless plugging for the continual personalization of evolved LLM at minimal cost. Our extensive experiments cover seven representative datasets, from easier question-answering tasks {BoolQ, SST2} to harder reasoning tasks {WinoGrande, GSM8K}, and models including {Mistral-7B, Llama2, Llama3.1, and Gemma2}, validating the portability of our designed model patches and showcasing the effectiveness of our proposed framework. For instance, PortLLM achieves comparable performance to LoRA fine-tuning with reductions of up to 12.2x in GPU memory usage. Finally, we provide theoretical justifications to understand the portability of our model update patches, which offers new insights into the theoretical dimension of LLMs' personalization.

cs.CL↗

arXiv2025

Make Optimization Once and for All with Fine-grained Guidance

Learning to Optimize (L2O) enhances optimization efficiency with integrated neural networks. L2O paradigms achieve great outcomes, e.g., refitting optimizer, generating unseen solutions iteratively or directly. However, conventional L2O methods require intricate design and rely on specific optimization processes, limiting scalability and generalization. Our analyses explore general framework for learning optimization, called Diff-L2O, focusing on augmenting sampled solutions from a wider view rather than local updates in real optimization process only. Meanwhile, we give the related generalization bound, showing that the sample diversity of Diff-L2O brings better performance. This bound can be simply applied to other fields, discussing diversity, mean-variance, and different tasks. Diff-L2O's strong compatibility is empirically verified with only minute-level training, comparing with other hour-levels.

cs.LG↗

arXiv2025

QuantMoE-Bench: Examining Post-Training Quantization for Mixture-of-Experts

Mixture-of-Experts (MoE) is a promising way to scale up the learning capacity of large language models. It increases the number of parameters while keeping FLOPs nearly constant during inference through sparse activation. Yet, it still suffers from significant memory overheads due to the vast parameter size, necessitating model compression techniques. Post-training quantization offers a powerful approach for model compression. Existing methods adopt a fixed quantization precision for the entire MoE model. This rigid setup can lead to suboptimal performance, without considering the inherent sparse structure. For example, MoE's sparse routing mechanism leads to different activation patterns, where shared experts are accessed by all tokens while token-conditioned experts are selectively activated. This activation disparity suggests different quantization requirements, with consistently activated shared experts potentially needing higher precision to maintain model quality. In this paper, we study a fine-grained precision setup for MoE quantization. We explore MoE structure-aware quantization heuristics, ranging from coarse (e.g., MoE layers) to fine granularity (e.g., linear layers). Our investigations reveal critical principles, where different MoE structures require varying numbers of bits for effective quantization. Conclusions are supported by extensive benchmarking across two representative MoE models and six tasks including commonsense reasoning and natural language understanding. We further show that an MoE quantized in a fined-grained mixed precision achieved state-of-the-art 65.35% performance on average compared to the baseline 64.30% (i.e., GPTQ). Moreover, based on the findings, we introduce novel data-driven techniques for optimizing bit allocation in MoE quantization, including the outlier-aware linear layer scorer and MoE block importance predictor.

cs.LG↗

arXiv2025

GroverGPT: A Large Language Model with 8 Billion Parameters for Quantum Searching

Quantum computing is an exciting non-Von Neumann paradigm, offering provable speedups over classical computing for specific problems. However, the practical limits of classical simulatability for quantum circuits remain unclear, especially with current noisy quantum devices. In this work, we explore the potential of leveraging Large Language Models (LLMs) to simulate the output of a quantum Turing machine using Grover's quantum circuits, known to provide quadratic speedups over classical counterparts. To this end, we developed GroverGPT, a specialized model based on LLaMA's 8-billion-parameter architecture, trained on over 15 trillion tokens. Unlike brute-force state-vector simulations, which demand substantial computational resources, GroverGPT employs pattern recognition to approximate quantum search algorithms without explicitly representing quantum states. Analyzing 97K quantum search instances, GroverGPT consistently outperformed OpenAI's GPT-4o (45\% accuracy), achieving nearly 100\% accuracy on 6- and 10-qubit datasets when trained on 4-qubit or larger datasets. It also demonstrated strong generalization, surpassing 95\% accuracy for systems with over 20 qubits when trained on 3- to 6-qubit data. Analysis indicates GroverGPT captures quantum features of Grover's search rather than classical patterns, supported by novel prompting strategies to enhance performance. Although accuracy declines with increasing system size, these findings offer insights into the practical boundaries of classical simulatability. This work suggests task-specific LLMs can surpass general-purpose models like GPT-4o in quantum algorithm learning and serve as powerful tools for advancing quantum research.

quant-ph↗

arXiv2024

Model-GLUE: Democratized LLM Scaling for A Large Model Zoo in the Wild

As Large Language Models (LLMs) excel across tasks and specialized domains, scaling LLMs based on existing models has garnered significant attention, which faces the challenge of decreasing performance when combining disparate models. Various techniques have been proposed for the aggregation of pre-trained LLMs, including model merging, Mixture-of-Experts, and stacking. Despite their merits, a comprehensive comparison and synergistic application of them to a diverse model zoo is yet to be adequately addressed. In light of this research gap, this paper introduces Model-GLUE, a holistic LLM scaling guideline. First, our work starts with a benchmarking of existing LLM scaling techniques, especially selective merging, and variants of mixture. Utilizing the insights from the benchmark results, we formulate an optimal strategy for the selection and aggregation of a heterogeneous model zoo characterizing different architectures and initialization.Our methodology involves the clustering of mergeable models and optimal merging strategy selection, and the integration of clusters through a model mixture. Finally, evidenced by our experiments on a diverse Llama-2-based model zoo, Model-GLUE shows an average performance enhancement of 5.61%, achieved without additional training. Codes are available at: https://github.com/Model-GLUE/Model-GLUE.

cs.LG↗

arXiv2024

Integrated magneto-photonic non-volatile multi-bit memory

We present an integrated magneto-photonic device for all-optical switching of non-volatile multi-bit spintronic memory. The bits are based on stand-alone magneto-tunnel junctions which are perpendicularly magnetized with all-optically switchable free layers, coupled onto photonic crystal nanobeam cavities on an indium phosphide based platform. This device enables switching of the magnetization state of the bits by locally increasing the power absorption of light at resonance with the cavity. We design an add/drop network of cavities to grant random access to multiple bits via a wavelength-division multiplexing scheme. Based on a three-dimensional finite-difference time-domain method, we numerically illustrate a compact device capable of switching and accessing 8 bits in different cavities with a 5 nm wavelength spacing in the conventional (C) telecommunication band. Our multi-bit device holds promise as a new paradigm for developing an ultrafast photonically-addressable spintronic memory and may also empower novel opportunities for photonically-driven spintronic-based neuromorphic computing.

physics.optics↗

arXiv2024

Enhancing Quantum Security over Federated Learning via Post-Quantum Cryptography

Federated learning (FL) has become one of the standard approaches for deploying machine learning models on edge devices, where private training data are distributed across clients, and a shared model is learned by aggregating locally computed updates from each client. While this paradigm enhances communication efficiency by only requiring updates at the end of each training epoch, the transmitted model updates remain vulnerable to malicious tampering, posing risks to the integrity of the global model. Although current digital signature algorithms can protect these communicated model updates, they fail to ensure quantum security in the era of large-scale quantum computing. Fortunately, various post-quantum cryptography algorithms have been developed to address this vulnerability, especially the three NIST-standardized algorithms - Dilithium, FALCON, and SPHINCS+. In this work, we empirically investigate the impact of these three NIST-standardized PQC algorithms for digital signatures within the FL procedure, covering a wide range of models, tasks, and FL settings. Our results indicate that Dilithium stands out as the most efficient PQC algorithm for digital signature in federated learning. Additionally, we offer an in-depth discussion of the implications of our findings and potential directions for future research.

quant-ph↗

arXiv2024

Picosecond all-optical switching of Co/Gd based synthetic ferrimagnets

Single pulse all-optical switching of magnetization (AOS) in Co/Gd based synthetic ferrimagnets carries promises for hybrid spintronic-photonic integration. A crucial next step progressing towards this vision is to gain insight into AOS and multi-domain state (MDS) behavior using longer pulses, which is compatible with state-of-the-art integrated photonics. In this work, we present our studies on the AOS and MDS of [Co/Gd]n (n = 1, 2) using ps optical pulses across a large composition range. We theoretically and experimentally show that a large Gd layer thickness can enhance the AOS energy efficiency and maximum pulse duration. We have identified two augmenting roles of Gd in extending the maximum pulse duration. On the inter-atomic level, we found that more Gd offers a prolonged angular momentum supply to Co. On the micromagnetic level, a higher Gd content brings the system to be closer to magnetic compensation in the equilibrized hot state, thereby reducing the driving force for thermally assisted nucleation of domain walls, combating the formation of a MDS. Our study presents a composition overview of AOS in [Co/Gd]n and offers useful physical insights regarding AOS fundamentals as well as the projected photonic integration.

cond-mat.mes-hall↗

arXiv2024

Revisiting Zeroth-Order Optimization for Memory-Efficient LLM Fine-Tuning: A Benchmark

In the evolving landscape of natural language processing (NLP), fine-tuning pre-trained Large Language Models (LLMs) with first-order (FO) optimizers like SGD and Adam has become standard. Yet, as LLMs grow {in size}, the substantial memory overhead from back-propagation (BP) for FO gradient computation presents a significant challenge. Addressing this issue is crucial, especially for applications like on-device training where memory efficiency is paramount. This paper proposes a shift towards BP-free, zeroth-order (ZO) optimization as a solution for reducing memory costs during LLM fine-tuning, building on the initial concept introduced by MeZO. Unlike traditional ZO-SGD methods, our work expands the exploration to a wider array of ZO optimization techniques, through a comprehensive, first-of-its-kind benchmarking study across five LLM families (Roberta, OPT, LLaMA, Vicuna, Mistral), three task complexities, and five fine-tuning schemes. Our study unveils previously overlooked optimization principles, highlighting the importance of task alignment, the role of the forward gradient method, and the balance between algorithm complexity and fine-tuning performance. We further introduce novel enhancements to ZO optimization, including block-wise descent, hybrid training, and gradient sparsity. Our study offers a promising direction for achieving further memory-efficient LLM fine-tuning. Codes to reproduce all our experiments are at https://github.com/ZO-Bench/ZO-LLM .

cs.LG↗

arXiv2024

Interlayer Dzyaloshinskii-Moriya interaction in synthetic ferrimagnets

The antisymmetric interlayer exchange interaction, i.e., interlayer Dzyaloshinskii-Moriya interaction (IL-DMI) has attracted significant interest since this long-range chiral spin interaction provides a new dimension for controlling spin textures and dynamics. However, the role of IL-DMI in the field induced and spin-orbit torque (SOT) induced switching of synthetic ferrimagnets (SFi) has not been uncovered. Here, we exploit interlayer chiral exchange bias fields in SFi to address both the sign and magnitude of the IL-DMI. Depending on the degree of imbalance between the two magnetic moments of the SFi, the amount of asymmetry, addressed via loop shifts of the hysteresis loops under an in-plane field reveals a unidirectional and chiral nature of the IL-DMI. The devices are then tested with SOT switching experiments and the process is examined via both transient state and steady state detection. In addition to field-free SOT switching, we find that the combination of IL-DMI and SOT give rise to multi-resistance states, which provides a possible direction for the future design of neuromorphic computing devices based on SOT. This work is a step towards characterizing and understanding the IL-DMI for spintronic applications.

physics.app-ph↗

arXiv2024

Merge, Then Compress: Demystify Efficient SMoE with Hints from Its Routing Policy

Sparsely activated Mixture-of-Experts (SMoE) has shown promise to scale up the learning capacity of neural networks, however, they have issues like (a) High Memory Usage, due to duplication of the network layers into multiple copies as experts; and (b) Redundancy in Experts, as common learning-based routing policies suffer from representational collapse. Therefore, vanilla SMoE models are memory inefficient and non-scalable, especially for resource-constrained downstream scenarios. In this paper, we ask: Can we craft a compact SMoE model by consolidating expert information? What is the best recipe to merge multiple experts into fewer but more knowledgeable experts? Our pilot investigation reveals that conventional model merging methods fail to be effective in such expert merging for SMoE. The potential reasons are: (1) redundant information overshadows critical experts; (2) appropriate neuron permutation for each expert is missing to bring all of them in alignment. To address this, we propose M-SMoE, which leverages routing statistics to guide expert merging. Specifically, it starts with neuron permutation alignment for experts; then, dominant experts and their "group members" are formed; lastly, every expert group is merged into a single expert by utilizing each expert's activation frequency as their weight for merging, thus diminishing the impact of insignificant experts. Moreover, we observed that our proposed merging promotes a low dimensionality in the merged expert's weight space, naturally paving the way for additional compression. Hence, our final method, MC-SMoE (i.e., Merge, then Compress SMoE), further decomposes the merged experts into low-rank and structural sparse alternatives. Extensive experiments across 8 benchmarks validate the effectiveness of MC-SMoE. For instance, our MC-SMoE achieves up to 80% memory and a 20% FLOPs reduction, with virtually no loss in performance.

cs.LG↗