Enmao Diao

2026

1. ICML

   OBCache: Optimal Brain KV Cache Pruning for Efficient Long-Context LLM Inference

   Yuzhe Gu, Xiyu Liang, Jiaojiao Zhao, and 1 more author

   In International Conference on Machine Learning (ICML), 2026

   Abs PDF Code

   Large language models (LLMs) with extended context windows enable powerful downstream applications but impose significant memory overhead, as caching all key-value (KV) states scales linearly with sequence length and batch size. Existing cache eviction methods address this by exploiting attention sparsity, yet they typically rank tokens heuristically using accumulated attention weights without considering their true impact on attention outputs. We propose Optimal Brain Cache (OBCache), a principled framework that formulates cache eviction as a layer-wise structured pruning problem. Building upon the Optimal Brain Damage (OBD) theory, OBCache quantifies token saliency by measuring the perturbation in attention outputs induced by pruning tokens, with closed-form scores derived for isolated keys, isolated values, and joint key-value pairs. Our scores account not only for attention weights but also for information from value states and attention outputs, thereby enhancing existing eviction strategies with output-aware signals. Experiments on LLaMA and Qwen models demonstrate that replacing the heuristic scores in existing works, which estimate token saliency across different query positions, with OBCache’s output-aware scores consistently improves long-context accuracy.

2. ICML

   A Kinetic-Energy Perspective of Flow Matching

   Ziyun Li, Huancheng Hu, Soon Hoe Lim, and 6 more authors

   In International Conference on Machine Learning (ICML), 2026

   Spotlight Abs PDF

   Spotlight

   Flow-based generative models can be viewed through a physics lens: sampling transports a particle from noise to data by integrating a time-varying velocity field, and each sample corresponds to a trajectory with its own dynamical effort. Motivated by classical mechanics, we introduce Kinetic Path Energy (KPE), an action-like, per-sample diagnostic that measures the accumulated kinetic effort along an Ordinary Differential Equation (ODE) trajectory. Empirically, KPE exhibits two robust correspondences: (i) higher KPE predicts stronger semantic fidelity; (ii) high-KPE trajectories terminate on low-density manifold frontiers. We further provide theoretical guarantees linking trajectory energy to data density. Paradoxically, this correlation is non-monotonic. At sufficiently high energy, generation can degenerate into memorization. Leveraging the closed-form of empirical flow matching, we show that extreme energies drive trajectories toward near-copies of training examples. This yields a Goldilocks principle and motivates Kinetic Trajectory Shaping (KTS), a training-free two-phase inference strategy that boosts early motion and enforces a late-time soft landing, reducing memorization and improving generation quality across benchmark tasks.

3. ACL

   From Local to Global: Revisiting Structured Pruning Paradigms for Large Language Models

   Ziyan Wang, Enmao Diao, Qi Le, and 6 more authors

   In Association for Computational Linguistics (ACL), 2026

   Abs PDF

   Structured pruning is a practical approach to deploying large language models (LLMs) efficiently, as it yields compact, hardware-friendly architectures. However, the dominant local paradigm is task-agnostic: by optimizing layer-wise reconstruction rather than task objectives, it tends to preserve perplexity or generic zero-shot behavior but fails to capitalize on modest task-specific calibration signals, often yielding limited downstream gains. We revisit global structured pruning and present GISP, Global Iterative Structured Pruning, a post-training method that removes attention heads and MLP channels using first-order, loss-based important scores aggregated at the structure level with block-wise normalization. Built on this global importance metric, GISP adopts an iterative schedule, rather than one-shot pruning, stabilizes accuracy at higher sparsity, and mitigates perplexity collapse without requiring intermediate fine-tuning. Importantly, the iterative pruning forms nested subnetworks that support a ”prune-once, deploy-many” workflow. Furthermore, GISP defines structural importance directly with respect to a target loss, making it easy to adapt pruning to task-specific objectives. In this work, we use perplexity for language modeling and a margin-based objective for decision-style tasks. Extensive experiments show that across Llama2-7B/13B, Llama3-8B, and Mistral-0.3-7B, GISP consistently lowers WikiText-2 perplexity and improves on downstream accuracy, with especially strong gains at 40-50% sparsity; on DeepSeek-R1-Distill-Llama-3-8B and Qwen3-8B with GSM8K, task-aligned calibration substantially boosts exact-match accuracy.

4. DAC

   Think Before You Prune: Self-Reflective Structured Pruning for Reasoning Language Models

   Ziyan Wang, Enmao Diao, Qi Le, and 5 more authors

   In Design Automation Conference (DAC), 2026

   Abs PDF

   Reasoning LLMs (RLMs) such as OpenAI o1, DeepSeek-R1, and Qwen3 deliver strong multi-step reasoning through chain-of-thought generation, but their large model sizes and lengthy decode-time outputs make them costly to deploy and unsuitable for resource-constrained settings. To reduce computing and memory cost, pruning offers a promising solution by removing unimportant parameters. However, despite their success on standard LLMs, existing pruning methods severely damage RLMs, as even moderate sparsity (e.g., 20%) can collapse accuracy and completely disrupt the model’s reasoning coherence. We begin by analyzing why existing pruning pipelines fail on reasoning LLMs and find that their brittleness largely stems from a mismatch between the calibration data, the pruning objective, and the model’s decode-time reasoning behavior. Our study further shows that the most reliable calibration signal comes not from human-written labels but from the model’s own self-generated reasoning traces, which more accurately reflect its inference distribution. Guided by these insights, we introduce RESP, a self-reflective structured pruning framework that aligns pruning decisions with the model’s reasoning dynamics through self-generated calibration, decode-only gradient-based importance estimation, and progressive regeneration that maintains calibration fidelity as sparsity increases. Experiments on Qwen3-8B demonstrate that RESP markedly outperforms existing structured pruning methods on both GSM8K and MathQA, preserving near-dense accuracy at 20-30% sparsity and substantially mitigating performance collapse at higher sparsity levels. At 40% sparsity, RESP attains 81.3% accuracy on GSM8K and 59.6% on MathQA, surpassing the strongest baselines by 66.87% and 47%, respectively.

5. ICLR

   Graph Tokenization for Bridging Graphs and Transformers

   Zeyuan Guo^*, Enmao Diao^*, Cheng Yang, and 1 more author

   In International Conference on Learning Representations (ICLR), 2026

   *Equal contribution

   Abs PDF Code

   The success of large pretrained Transformers is closely tied to tokenizers, which convert raw input into discrete symbols. Extending these models to graph-structured data remains a significant challenge. In this work, we introduce a graph tokenization framework that generates sequential representations of graphs by combining reversible graph serialization, which preserves graph information, with Byte Pair Encoding (BPE), a widely adopted tokenizer in large language models (LLMs). To better capture structural information, the graph serialization process is guided by global statistics of graph substructures, ensuring that frequently occurring substructures appear more often in the sequence and can be merged by BPE into meaningful tokens. Empirical results demonstrate that the proposed tokenizer enables Transformers such as BERT to be directly applied to graph benchmarks without architectural modifications. The proposed approach achieves state-of-the-art results on 14 benchmark datasets and frequently outperforms both graph neural networks and specialized graph transformers. This work bridges the gap between graph-structured data and the ecosystem of sequence models.

2025

1. arXiv

   Toward Unifying Group Fairness Evaluation from a Sparsity Perspective

   Zhecheng Sheng, Jiawei Zhang, and Enmao Diao

   arXiv, 2025

   Abs PDF Code

   Ensuring algorithmic fairness remains a significant challenge in machine learning, particularly as models are increasingly applied across diverse domains. While numerous fairness criteria exist, they often lack generalizability across different machine learning problems. This paper examines the connections and differences among various sparsity measures in promoting fairness and proposes a unified sparsity-based framework for evaluating algorithmic fairness. The framework aligns with existing fairness criteria and demonstrates broad applicability to a wide range of machine learning tasks. We demonstrate the effectiveness of the proposed framework as an evaluation metric through extensive experiments on a variety of datasets and bias mitigation methods. This work provides a novel perspective to algorithmic fairness by framing it through the lens of sparsity and social equity, offering potential for broader impact on fairness research and applications.

2. NeurIPS

   Beyond Expectations: Quantile-Guided Alignment for Risk-Calibrated Language Models

   Xinran Wang, Jin Du, Azal Ahmad Khan, and 5 more authors

   In Advances in Neural Information Processing Systems (NeurIPS), 2025

   Spotlight Abs PDF Code

   Spotlight

   Large language models can generate rare but catastrophic outputs, such as harmful conversations or insecure code. Existing Reinforcement Learning from Human Feedback (RLHF) typically maximizes average reward, leaving high-risk tail events insufficiently controlled. We introduce Quantile‑Guided Alignment (QA), a framework that allows users to specify desired improvements at any quantile—individually or across multiple reward dimensions—thus shifting the distribution of outputs with finer control toward safer, more desirable outcomes. The method extends standard RLHF via an augmented reward formulation that enforces quantile constraints. Experiments on conversation and code‐generation tasks show that quantile alignment significantly enhances quality at targeted tails while maintaining overall performance. The results position QA as a principled route to risk‑calibrated language models with tail‑focused alignment.

3. ICLR

   MAP: Multi-Human-Value Alignment Palette

   Xinran Wang, Qi Le, Ammar Ahmed, and 5 more authors

   In International Conference on Learning Representations (ICLR), 2025

   Oral Abs PDF Code

   Oral

   Ensuring that generative AI systems align with human values is essential but challenging, especially when considering multiple human values and their potential trade-offs. Since human values can be personalized and dynamically change over time, the desirable levels of value alignment vary across different ethnic groups, industry sectors, and user cohorts. Within existing frameworks, it is hard to define human values and align AI systems accordingly across different directions simultaneously, such as harmlessness, helpfulness, and positiveness. To address this, we develop a novel, first-principle approach called Multi-Human-Value Alignment Palette (MAP), which navigates the alignment across multiple human values in a structured and reliable way. MAP formulates the alignment problem as an optimization task with user-defined constraints, which define human value targets. It can be efficiently solved via a primal-dual approach, which determines whether a user-defined alignment target is achievable and how to achieve it. We conduct a detailed theoretical analysis of MAP by quantifying the trade-offs between values, the sensitivity to constraints, the fundamental connection between multi-value alignment and sequential alignment, and proving that linear weighted rewards are sufficient for multi-value alignment. Extensive experiments demonstrate MAP’s ability to align multiple values in a principled manner while delivering strong empirical performance across various tasks.

4. ICLR

   Probe Pruning: Accelerating LLMs through Dynamic Pruning via Model-Probing

   Qi Le, Enmao Diao, Ziyan Wang, and 4 more authors

   In International Conference on Learning Representations (ICLR), 2025

   Abs PDF Code

   We introduce Probe Pruning (PP), a novel framework for online, dynamic, structured pruning of Large Language Models (LLMs) applied in a batch-wise manner. PP leverages the insight that not all samples and tokens contribute equally to the model’s output, and probing a small portion of each batch effectively identifies crucial weights, enabling tailored dynamic pruning for different batches. It comprises three main stages: probing, history-informed pruning, and full inference. In the probing stage, PP selects a small yet crucial set of hidden states, based on residual importance, to run a few model layers ahead. During the history-informed pruning stage, PP strategically integrates the probing states with historical states. Subsequently, it structurally prunes weights based on the integrated states and the PP importance score, a metric developed specifically to assess the importance of each weight channel in maintaining performance. In the final stage, full inference is conducted on the remaining weights. A major advantage of PP is its compatibility with existing models, as it operates without requiring additional neural network modules or fine-tuning. Comprehensive evaluations of PP on LLaMA-2/3 and OPT models reveal that even minimal probing-using just 1.5% of FLOPs-can substantially enhance the efficiency of structured pruning of LLMs. For instance, when evaluated on LLaMA-2-7B with WikiText2, PP achieves a 2.56 times lower ratio of performance degradation per unit of runtime reduction compared to the state-of-the-art method at a 40% pruning ratio.

5. NAACL

   AID: Adaptive Integration of Detectors for Safe AI with Language Models

   Xinran Wang, Enmao Diao, Qi Le, and 2 more authors

   In Nations of the Americas Chapter of the ACL (NAACL), 2025

   Abs PDF

   As Large Language Models (LLMs) increasingly influence content generation across diverse platforms, there is a heightened urgency to regulate their outputs to ensure safe usage. However, defining “safety” is complex, given that entities across domains may interpret it through varied lenses and develop safety detectors—models trained to identify specific unsafe content based on predefined criteria. To address this complexity, we introduce the approach of Adaptive Integration of Detectors (AID) to orchestrate the strengths of multiple pretrained detectors to ensure comprehensive effectiveness in diverse scenarios. AID employs a Mixture-of-Experts (MoE) framework, wherein it dynamically assigns and learns data-adaptive weights for each detector using domain-specific annotated data and LLM-extracted features. We provide theoretical insights into why MoE can be effective by showing its optimality in a Neyman-Pearson setting. Our experimental studies using various detection tasks curated from benchmark datasets demonstrate AID’s ability to synergistically combine the unique capabilities of individual detectors. For example, it is observed that AID can improve the area under the curve (AUC) by an absolute value of 0.07 to 0.21, with a median of 0.12, compared to the best individual detectors developed for specific safety aspects. The improvement is particularly significant for complex detection tasks that mix different unsafe data sources.

2024

1. IEEE BigData

   DynamicFL: Federated Learning with Dynamic Communication Resource Allocation

   Qi Le, Enmao Diao, Xinran Wang, and 4 more authors

   In IEEE International Conference on Big Data (BigData), 2024

   Best Student Paper Abs PDF Code

   Best Student Paper

   Federated Learning (FL) is a collaborative machine learning framework that allows multiple users to train models utilizing their local data in a distributed manner. However, considerable statistical heterogeneity in local data across devices often leads to suboptimal model performance compared with independently and identically distributed (IID) data scenarios. In this paper, we introduce DynamicFL, a new FL framework that investigates the trade-offs between global model performance and communication costs for two widely adopted FL methods: Federated Stochastic Gradient Descent (FedSGD) and Federated Averaging (FedAvg). Our approach allocates diverse communication resources to clients based on their data statistical heterogeneity, considering communication resource constraints, and attains substantial performance enhancements compared to uniform communication resource allocation. Notably, our method bridges the gap between FedSGD and FedAvg, providing a flexible framework leveraging communication heterogeneity to address statistical heterogeneity in FL. Through extensive experiments, we demonstrate that DynamicFL surpasses current state-of-the-art methods with up to a 10% increase in model accuracy, demonstrating its adaptability and effectiveness in tackling data statistical heterogeneity challenges.

2. EMNLP

   ESC: Efficient Speech Coding with Cross-Scale Residual Vector Quantized Transformers

   Yuzhe Gu, and Enmao Diao

   In Empirical Methods in Natural Language Processing (EMNLP), 2024

   Abs PDF Code

   Neural speech codecs aim to compress input signals into minimal bits while maintaining content quality in a low-latency manner. However, existing neural codecs often trade model complexity for reconstruction performance. These codecs primarily use convolutional blocks for feature transformation, which are not inherently suited for capturing the local redundancies in speech signals. To compensate, they require either adversarial discriminators or a large number of model parameters to enhance audio quality. In response to these challenges, we introduce the Efficient Speech Codec (ESC), a lightweight, parameter-efficient speech codec based on a cross-scale residual vector quantization scheme and transformers. Our model employs mirrored hierarchical window transformer blocks and performs step-wise decoding from coarse-to-fine feature representations. To enhance bitrate efficiency, we propose a novel combination of vector quantization techniques along with a pre-training paradigm. Extensive experiments demonstrate that ESC can achieve high-fidelity speech reconstruction with significantly lower model complexity, making it a promising alternative to existing convolutional audio codecs

3. IEEE Access

   Large Deviation Analysis of Score-based Hypothesis Testing

   Enmao Diao, Taposh Banerjee, and Vahid Tarokh

   IEEE Access, 2024

   Abs PDF Code

   Score-based statistical models play an important role in modern machine learning, statistics, and signal processing. For hypothesis testing, a score-based hypothesis test is proposed in \citewu2022score. We analyze the performance of this score-based hypothesis testing procedure and derive upper bounds on the probabilities of its Type I and II errors. We prove that the exponents of our error bounds are asymptotically (in the number of samples) tight for the case of simple null and alternative hypotheses. We calculate these error exponents explicitly in specific cases and provide numerical studies for various other scenarios of interest.

4. RadarConf

   A PixelCNN Based Method for Rough Surface Clutter Reduction in GPR B-scan Images

   Yan Zhang, Enmao Diao, Dryver Huston, and 1 more author

   In IEEE Radar Conference (RadarConf), 2024

   Abs PDF

   Reducing the rough surface clutter in a Ground Penetrating Radar (GPR) B-scan image is essential for detecting shallow buried targets and for improving the performance of subsequent image processing algorithms. Challenges involved in this problem lie in the difficulties with modeling the variations of the rough surface profile. In this paper, we propose a novel PixelCNN based method for reducing rough surface clutter in GPR B-scan images. In the proposed method, the rough surface region in a B-scan image is split into small patches which are used to train a PixelCNN model. Given an input patch, the trained PixelCNN model can output the probability that the input patch comes from the rough surface region. For reconstructing a clutter reduced B-scan, the entire B-scan image is split into patches and each patch is input into the trained PixelCNN model to get its corresponding probability. Negative log-probabilities are then utilized as scores to suppress the rough surface region and to enhance the target profile in the B-scan image. To demonstrate the effectiveness of the proposed method, we test it on four simulation B-scan images and two real-world B-scan images. The proposed method is compared with traditional subspace projection methods. The results indicate that the proposed method outperforms traditional subspace projection methods.

5. TGRS

   A Data Efficient Deep Learning Method for Rough Surface Clutter Reduction in GPR Images

   Yan Zhang, Enmao Diao, Dryver Huston, and 1 more author

   IEEE Transactions on Geoscience and Remote Sensing (TGRS), 2024

   Abs PDF

   In ground penetrating radar (GPR) B-scan images, ground surface clutter is the main source of interference, often obscuring or distorting subsurface target signals. We propose a deep autoencoder-based method to mitigate rough surface clutter in GPR images by treating it as an anomaly detection problem. First, the rough surface region in a B-scan image is partitioned into small patches, which act as the training dataset for the deep autoencoder. Through the training process, the autoencoder learns and captures the patterns associated with the rough surface patches. Following training, the entire B-scan image is divided into small patches of the same size as the training patches, and each of them is fed into the autoencoder to compute an anomaly score. To reconstruct a clutter-reduced B-scan image, we employ a weighted sum approach to aggregate all patches based on their anomaly scores. We evaluate our method against conventional subspace projection techniques using simulated and field-collected B-scans. The results clearly indicate that our approach surpasses these subspace methods. Furthermore, we employ t-distributed stochastic neighbor embedding (t-SNE) analysis to gain deeper insights into our method’s effectiveness in reducing rough surface clutter. The outcomes of this analysis reinforce the practical viability of our approach for GPR image processing.

2023

1. TIT

   Quickest Change Detection for Unnormalized Statistical Models

   Suya Wu, Enmao Diao, Taposh Banerjee, and 2 more authors

   IEEE Transactions on Information Theory (TIT), 2023

   Abs PDF Code

   Classical quickest change detection algorithms require modeling pre-change and post-change distributions. Such an approach may not be feasible for various machine learning models because of the complexity of computing the explicit distributions. Additionally, these methods may suffer from a lack of robustness to model mismatch and noise. This paper develops a new variant of the classical Cumulative Sum (CUSUM) algorithm for the quickest change detection. This variant is based on Fisher divergence and the Hyvärinen score and is called the Hyvärinen score-based CUSUM (SCUSUM) algorithm. The SCUSUM algorithm allows the applications of change detection for unnormalized statistical models, i.e., models for which the probability density function contains an unknown normalization constant. The asymptotic optimality of the proposed algorithm is investigated by deriving expressions for average detection delay and the mean running time to a false alarm. Numerical results are provided to demonstrate the performance of the proposed algorithm.

2. Thesis

   Efficient and Collaborative Methods for Distributed Machine Learning

   Enmao Diao

   Duke University, 2023

   Abs PDF

   In recent years, there has been a significant expansion in the scale and complexity of neural networks. This has resulted in significant demand for data, computation, and energy resources. In this light, it is crucial to enhance and optimize the efficiency of these ML models and algorithms. Additionally, the rise in computational capabilities of modern devices has prompted a shift towards distributed systems that enable localized data storage and model training. While this evolution promises substantial potential, it introduces a series of challenges. Such challenges encompass addressing the heterogeneity across systems, data, models, and supervision, balancing the trade-off among communication, computation, and performance, as well as building a community of shared interest to encourage collaboration in the emerging era of Artificial General Intelligence (AGI).In this dissertation, we contribute to the establishment of a theoretically justified, methodologically comprehensive, and universally applicable Efficient and Collaborative Distributed Machine Learning framework. Specifically, in Part I, we contribute to methodologies for Efficient Machine Learning including for both learning and inference. In this direction, we propose a parameter-efficient model, namely Restricted Recurrent Neural Networks (RRNN), that leverage the recurrent structures of RNNs using weight sharing in order to improve learning efficiency. We also introduce an optimal measure of vector sparsity named the PQ Index (PQI), and postulate a hypothesis connecting this sparsity measure and compressibility of neural networks. Based on this, we propose a Sparsity-informed Adaptive Pruning (SAP) algorithm. This algorithm adaptively determines the pruning ratio to enhance inference efficiency. In Part II, we address both efficiency and collaboration in Distributed Machine Learning. We introduce Distributed Recurrent Autoencoders for Scalable Image Compression (DRASIC), a data-driven Distributed Source Coding framework that can compress heterogeneous data in a scalable and distributed manner. We then propose Heterogeneous Federated Learning (HeteroFL), demonstrating the feasibility of training localized heterogeneous models to create a global inference model. Subsequently, we propose a new Federated Learning (FL) framework, namely SemiFL, to tackle Semi-Supervised Federated Learning (SSFL) for clients with unlabeled data. This method performs comparably with state-of-the-art centralized Semi-Supervised Learning (SSL), and fully supervised FL techniques. Finally, we propose Gradient Assisted Learning (GAL) in order to enable collaborations among multiple organizations without sharing data, models, and objective functions. This method significantly out-performs local learning baselines and achieves near-oracle performance. In Part III, we develop collaborative applications for building a community of shared interest. We apply SemiFL to Keyword Spotting (KWS), a technique widely used in virtual assistants. Numerical experiments demonstrate that one can train models from the scratch, or transfer from pre-trained models in order to leverage heterogeneous unlabeled on-device data, using only a small amount of labeled data from the server. Finally, we propose a Decentralized Multi-Target Cross-Domain Recommendation (DMTCDR) which enhances the recommendation performance of decentralized organizations without compromising data privacy or model confidentiality.

3. UAI

   Robust Quickest Change Detection for Unnormalized Models

   Suya Wu, Enmao Diao, Jie Ding, and 2 more authors

   In Uncertainty in Artificial Intelligence (UAI), 2023

   Abs PDF

   Detecting an abrupt and persistent change in the underlying distribution of online data streams is an important problem in many applications. This paper proposes a new robust score-based algorithm called RSCUSUM, which can be applied to unnormalized models and addresses the issue of unknown post-change distributions. RSCUSUM replaces the Kullback-Leibler divergence with the Fisher divergence between pre-and post-change distributions for computational efficiency in unnormalized statistical models and introduces a notion of the “least favorable” distribution for robust change detection. The algorithm and its theoretical analysis are demonstrated through simulation studies.

4. ICLR

   Pruning Deep Neural Networks from a Sparsity Perspective

   Enmao Diao^*, Ganghua Wang^*, Jiawei Zhang, and 3 more authors

   In International Conference on Learning Representations (ICLR), 2023

   *Equal contribution

   Abs PDF Code

   In recent years, deep network pruning has attracted significant attention in order to enable the rapid deployment of AI into small devices with computation and memory constraints. Pruning is often achieved by dropping redundant weights, neurons, or layers of a deep network while attempting to retain a comparable test performance. Many deep pruning algorithms have been proposed with impressive empirical success. However, existing approaches lack a quantifiable measure to estimate the compressibility of a sub-network during each pruning iteration and thus may under-prune or over-prune the model. In this work, we propose PQ Index (PQI) to measure the potential compressibility of deep neural networks and use this to develop a Sparsity-informed Adaptive Pruning (SAP) algorithm. Our extensive experiments corroborate the hypothesis that for a generic pruning procedure, PQI decreases first when a large model is being effectively regularized and then increases when its compressibility reaches a limit that appears to correspond to the beginning of underfitting. Subsequently, PQI decreases again when the model collapse and significant deterioration in the performance of the model start to occur. Additionally, our experiments demonstrate that the proposed adaptive pruning algorithm with proper choice of hyper-parameters is superior to the iterative pruning algorithms such as the lottery ticket-based pruning methods, in terms of both compression efficiency and robustness.

5. AISTATS

   Score-based Quickest Change Detection for Unnormalized Models

   Suya Wu, Enmao Diao, Taposh Banerjee, and 2 more authors

   In International Conference on Artificial Intelligence and Statistics (AISTATS), 2023

   Abs PDF Code

   Classical change detection algorithms typically require modeling pre-change and post-change distributions. The calculations may not be feasible for various machine learning models because of the complexity of computing the partition functions and normalized distributions. Additionally, these methods may suffer from a lack of robustness to model mismatch and noise. In this paper, we develop a new variant of the classical Cumulative Sum (CUSUM) change detection, namely Score-based CUSUM (SCUSUM), based on Fisher divergence and the Hyvärinen score. Our method allows the applications of the quickest change detection for unnormalized distributions. We provide a theoretical analysis of the detection delay given the constraints on false alarms. We prove the asymptotic optimality of the proposed method in some particular cases. We also provide numerical experiments to demonstrate our method’s computation, performance, and robustness advantages.

2022

1. NeurIPS

   GAL: Gradient Assisted Learning for Decentralized Multi-organization Collaborations

   Enmao Diao, Jie Ding, and Vahid Tarokh

   In Advances in Neural Information Processing Systems (NeurIPS), 2022

   Abs PDF Code

   Collaborations among multiple organizations, such as financial institutions, medical centers, and retail markets in decentralized settings are crucial to providing improved service and performance. However, the underlying organizations may have little interest in sharing their local data, models, and objective functions. These requirements have created new challenges for multi-organization collaboration. In this work, we propose Gradient Assisted Learning (GAL), a new method for multiple organizations to assist each other in supervised learning tasks without sharing local data, models, and objective functions. In this framework, all participants collaboratively optimize the aggregate of local loss functions, and each participant autonomously builds its own model by iteratively fitting the gradients of the overarching objective function. We also provide asymptotic convergence analysis and practical case studies of GAL. Experimental studies demonstrate that GAL can achieve performance close to centralized learning when all data, models, and objective functions are fully disclosed.

2. NeurIPS

   SemiFL: Semi-Supervised Federated Learning for Unlabeled Clients with Alternate Training

   Enmao Diao, Jie Ding, and Vahid Tarokh

   In Advances in Neural Information Processing Systems (NeurIPS), 2022

   Abs PDF Code

   Federated Learning allows the training of machine learning models by using the computation and private data resources of many distributed clients. Most existing results on Federated Learning (FL) assume the clients have ground-truth labels. However, in many practical scenarios, clients may be unable to label task-specific data due to a lack of expertise or resource. We propose SemiFL to address the problem of combining communication-efficient FL such as FedAvg with Semi-Supervised Learning (SSL). In SemiFL, clients have completely unlabeled data and can train multiple local epochs to reduce communication costs, while the server has a small amount of labeled data. We provide a theoretical understanding of the success of data augmentation-based SSL methods to illustrate the bottleneck of a vanilla combination of communication-efficient FL with SSL. To address this issue, we propose alternate training to ’fine-tune global model with labeled data’ and ’generate pseudo-labels with the global model.’ We conduct extensive experiments and demonstrate that our approach significantly improves the performance of a labeled server with unlabeled clients training with multiple local epochs. Moreover, our method outperforms many existing SSFL baselines and performs competitively with the state-of-the-art FL and SSL results.

2021

1. ICLR

   HeteroFL: Computation and Communication Efficient Federated Learning for Heterogeneous Clients

   Enmao Diao, Jie Ding, and Vahid Tarokh

   In International Conference on Learning Representations (ICLR), 2021

   Abs PDF Code

   Federated Learning (FL) is a method of training machine learning models on private data distributed over a large number of possibly heterogeneous clients such as mobile phones and IoT devices. In this work, we propose a new federated learning framework named HeteroFL to address heterogeneous clients equipped with very different computation and communication capabilities. Our solution can enable the training of heterogeneous local models with varying computation complexities and still produce a single global inference model. For the first time, our method challenges the underlying assumption of existing work that local models have to share the same architecture as the global model. We demonstrate several strategies to enhance FL training and conduct extensive empirical evaluations, including five computation complexity levels of three model architecture on three datasets. We show that adaptively distributing subnetworks according to clients’ capabilities is both computation and communication efficient.

2020

1. ICASSP

   Speech Emotion Recognition with Dual-Sequence LSTM Architecture

   Jianyou Wang, Michael Xue, Ryan Culhane, and 3 more authors

   In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2020

   Abs PDF Code

   Speech Emotion Recognition (SER) has emerged as a critical component of the next generation of human-machine interfacing technologies. In this work, we propose a new dual-level model that predicts emotions based on both MFCC features and mel-spectrograms produced from raw audio signals. Each utterance is preprocessed into MFCC features and two mel-spectrograms at different time-frequency resolutions. A standard LSTM processes the MFCC features, while a novel LSTM architecture, denoted as Dual-Sequence LSTM (DS-LSTM), processes the two mel-spectrograms simultaneously. The outputs are later averaged to produce a final classification of the utterance. Our proposed model achieves, on average, a weighted accuracy of 72.7% and an unweighted accuracy of 73.3% - a 6% improvement over current state-of-the-art unimodal models - and is comparable with multimodal models that leverage textual information as well as audio signals.

2. DCC

   DRASIC: Distributed Recurrent Autoencoder for Scalable Image Compression

   Enmao Diao, Jie Ding, and Vahid Tarokh

   In Data Compression Conference (DCC), 2020

   Abs PDF Code

   We propose a new architecture for distributed image compression from a group of distributed data sources. The work is motivated by practical needs of data-driven codec design, low power consumption, robustness, and data privacy. The proposed architecture, which we refer to as Distributed Recurrent Autoencoder for Scalable Image Compression (DRASIC), is able to train distributed encoders and one joint decoder on correlated data sources. Its compression capability is much better than the method of training codecs separately. Meanwhile, the performance of our distributed system with 10 distributed sources is only within 2 dB peak signal-to-noise ratio (PSNR) of the performance of a single codec trained with all data sources. We experiment distributed sources with different correlations and show how our data-driven methodology well matches the Slepian-Wolf Theorem in Distributed Source Coding (DSC). To the best of our knowledge, this is the first data-driven DSC framework for general distributed code design with deep learning.

2019

1. IEEE Big Data

   Restricted Recurrent Neural Networks

   Enmao Diao, Jie Ding, and Vahid Tarokh

   In IEEE International Conference on Big Data (Big Data), 2019

   Student Travel Award Abs PDF Code

   Student Travel Award

   Recurrent Neural Network (RNN) and its variations such as Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU), have become standard building blocks for learning online data of sequential nature in many research areas, including natural language processing and speech data analysis. In this paper, we present a new methodology to significantly reduce the number of parameters in RNNs while maintaining performance that is comparable or even better than classical RNNs. The new proposal, referred to as Restricted Recurrent Neural Network (RRNN), restricts the weight matrices corresponding to the input data and hidden states at each time step to share a large proportion of parameters. The new architecture can be regarded as a compression of its classical counterpart, but it does not require pre-training or sophisticated parameter fine-tuning, both of which are major issues in most existing compression techniques. Experiments on natural language modeling show that compared with its classical counterpart, the restricted recurrent architecture generally produces comparable results at about 50% compression rate. In particular, the Restricted LSTM can outperform classical RNN with even less number of parameters.