While 3D instance segmentation (3DIS) has advanced significantly, existing methods typically assume that all object classes are known in advance and are uniformly distributed. However, this assumption is unrealistic in dynamic, real-world environments where new classes emerge gradually and exhibit natural imbalance. Although some approaches have addressed class emergence, they often overlook class imbalance, resulting in suboptimal performance -- particularly on rare categories. To tackle this challenge, we propose CLIMB-3D, a unified framework for \textbf{CL}ass-incremental \textbf{Imb}alance-aware \textbf{3D}IS. Building upon established exemplar replay (ER) strategies, we show that ER alone is insufficient to achieve robust performance under constrained memory conditions. To mitigate this, we introduce a novel pseudo-label generator (PLG) that extends supervision to previously learned categories by leveraging predictions from a frozen prior model. Despite its promise, PLG tends to bias towards frequent classes. Therefore, we propose a class-balanced re-weighting (CBR) scheme, that estimates object frequencies from pseudo-labels and dynamically adjusts training bias -- without requiring access to past data. We design and evaluate three incremental scenarios for 3DIS on the challenging ScanNet200 dataset, and additionally on semantic segmentation on ScanNetV2. Our approach achieves state-of-the-art results, surpassing prior work by up to 16.76\% mAP for instance segmentation and approximately 30\% mIoU for semantic segmentation, demonstrating strong generalization across both frequent and rare classes.
With the rapid development of artificial intelligence (AI), especially in the medical field, the need for its explainability has grown. In medical image analysis, a high degree of transparency and model interpretability can help clinicians better understand and trust the decision-making process of AI models. In this study, we propose a Knowledge Distillation (KD)-based approach that aims to enhance the transparency of the AI model in medical image analysis. The initial step is to use traditional CNN to obtain a teacher model and then use KD to simplify the CNN architecture, retain most of the features of the data set, and reduce the number of network layers. It also uses the feature map of the student model to perform hierarchical analysis to identify key features and decision-making processes. This leads to intuitive visual explanations. We selected three public medical data sets (brain tumor, eye disease, and Alzheimer's disease) to test our method. It shows that even when the number of layers is reduced, our model provides a remarkable result in the test set and reduces the time required for the interpretability analysis.
Change detection (CD) in remote sensing images plays a vital role in Earth observation. However, the scarcity of high-resolution, comprehensive open-source datasets and the difficulty in achieving robust performance across varying change types remain major challenges. To address these issues, we introduce JL1-CD, a large-scale, sub-meter CD dataset consisting of 5,000 image pairs. We further propose a novel Origin-Partition (O-P) strategy and integrate it into a Multi-Teacher Knowledge Distillation (MTKD) framework to enhance CD performance. The O-P strategy partitions the training set by Change Area Ratio (CAR) and trains specialized teacher models on each subset. The MTKD framework then distills complementary knowledge from these teachers into a single student model, enabling improved detection results across diverse CAR scenarios without additional inference cost. Our MTKD approach demonstrated strong performance in the 2024 "Jilin-1'' Cup challenge, ranking first in the preliminary and second in the final rounds. Extensive experiments on the JL1-CD and SYSU-CD datasets show that the MTKD framework consistently improves the performance of CD models with various network architectures and parameter sizes, establishing new state-of-the-art results. Code and dataset are available at https://anonymous.4open.science/r/MTKD-A-84B8.
As synthetic data becomes increasingly popular in machine learning tasks, numerous methods--without formal differential privacy guarantees--use synthetic data for training. These methods often claim, either explicitly or implicitly, to protect the privacy of the original training data. In this work, we explore four different training paradigms: coreset selection, dataset distillation, data-free knowledge distillation, and synthetic data generated from diffusion models. While all these methods utilize synthetic data for training, they lead to vastly different conclusions regarding privacy preservation. We caution that empirical approaches to preserving data privacy require careful and rigorous evaluation; otherwise, they risk providing a false sense of privacy.
Recent 2D CNN-based domain adaptation approaches struggle with long-range dependencies due to limited receptive fields, making it difficult to adapt to target domains with significant spatial distribution changes. While transformer-based domain adaptation methods better capture distant relationships through self-attention mechanisms that facilitate more effective cross-domain feature alignment, their quadratic computational complexity makes practical deployment challenging for object detection tasks across diverse domains. Inspired by the global modeling and linear computation complexity of the Mamba architecture, we present the first domain-adaptive Mamba-based one-stage object detection model, termed DA-Mamba. Specifically, we combine Mamba's efficient state-space modeling with attention mechanisms to address domain-specific spatial and channel-wise variations. Our design leverages domain-adaptive spatial and channel-wise scanning within the Mamba block to extract highly transferable representations for efficient sequential processing, while cross-attention modules generate long-range, mixed-domain spatial features to enable robust soft alignment across domains. Besides, motivated by the observation that hybrid architectures introduce feature noise in domain adaptation tasks, we propose an entropy-based knowledge distillation framework with margin ReLU, which adaptively refines multi-level representations by suppressing irrelevant activations and aligning uncertainty across source and target domains. Finally, to prevent overfitting caused by the mixed-up features generated through cross-attention mechanisms, we propose entropy-driven gating attention with random perturbations that simultaneously refine target features and enhance model generalization.
The widespread deployment of Large Language Models (LLMs) is hindered by the high computational demands, making knowledge distillation (KD) crucial for developing compact smaller ones. However, the conventional KD methods endure the distribution mismatch issue between the teacher and student models, leading to the poor performance of distillation. For instance, the widely-used KL-based methods suffer the mode-averaging and mode-collapsing problems, since the mismatched probabitliy distribution between both models. Previous studies mainly optimize this issue via different distance calculations towards the distribution of both models. Unfortunately, the distribution mismatch issue still exists in the early stage of the distillation. Hence, to reduce the impact of distribution mismatch, we propose a simple yet efficient method, named Warmup-Distill, which aligns the distillation of the student to that of the teacher in advance of distillation. Specifically, we first detect the distribution of the student model in practical scenarios with its internal knowledge, and then modify the knowledge with low probability via the teacher as the checker. Consequently, Warmup-Distill aligns the internal student's knowledge to that of the teacher, which expands the distribution of the student with the teacher's, and assists the student model to learn better in the subsequent distillation. Experiments on the seven benchmarks demonstrate that Warmup-Distill could provide a warmup student more suitable for distillation, which outperforms the vanilla student by as least +0.4 averaged score among all benchmarks. Noteably, with the assistance of Warmup-Distill, the distillation on the math task could yield a further improvement, at most +1.9% accuracy.
The radioactive nature of Large Language Model (LLM) watermarking enables the detection of watermarks inherited by student models when trained on the outputs of watermarked teacher models, making it a promising tool for preventing unauthorized knowledge distillation. However, the robustness of watermark radioactivity against adversarial actors remains largely unexplored. In this paper, we investigate whether student models can acquire the capabilities of teacher models through knowledge distillation while avoiding watermark inheritance. We propose two categories of watermark removal approaches: pre-distillation removal through untargeted and targeted training data paraphrasing (UP and TP), and post-distillation removal through inference-time watermark neutralization (WN). Extensive experiments across multiple model pairs, watermarking schemes and hyper-parameter settings demonstrate that both TP and WN thoroughly eliminate inherited watermarks, with WN achieving this while maintaining knowledge transfer efficiency and low computational overhead. Given the ongoing deployment of watermarking techniques in production LLMs, these findings emphasize the urgent need for more robust defense strategies. Our code is available at https://github.com/THU-BPM/Watermark-Radioactivity-Attack.
Knowledge Distillation (KD) has emerged as a prominent technique for model compression. However, conventional KD approaches primarily focus on homogeneous architectures with identical tokenizers, constraining their applicability in cross-architecture scenarios. As for the cross-tokenizer KD, the differences in the tokenizers give rise to two fundamental challenges: (1) sequence misalignment caused by divergent tokenization strategies, and (2) mismatched vocabulary size and composition. While existing probability-matching methods attempt to address these issues, their efficacy remains limited due to suboptimal alignment in both the sequence and vocabulary aspects. To overcome these limitations, we propose Contextual Dynamic Mapping (CDM), a novel cross-tokenizer distillation framework that employs contextual information to enhance sequence alignment precision and dynamically improves vocabulary mapping. We evaluated the effectiveness of our approach across five advanced and widely-used model families (i.e, LLama3, Phi3, Gemma2, OPT and Qwen2), which were configured into three distinct teacher-student pairs. Our method shows significant advantages over existing cross-tokenizer distillation baselines across diverse benchmarks, including instruction-following, code generation and math. Notably, our analysis reveals that combining conventional same-tokenizer distillation and cross-tokenizer distillation through CDM yields further performance improvements. The code is available at https://github.com/pppa2019/ContexualDynamicMapping
Social recommendation, which incorporates social connections into recommender systems, has proven effective in improving recommendation accuracy. However, beyond accuracy, diversity is also crucial for enhancing user engagement. Despite its importance, the impact of social recommendation models on diversity remains largely unexplored. In this study, we systematically examine the dual performance of existing social recommendation algorithms in terms of both accuracy and diversity. Our empirical analysis reveals a concerning trend: while social recommendation models enhance accuracy, they often reduce diversity. To address this issue, we propose Diversified Social Recommendation (DivSR), a novel approach that employs relational knowledge distillation to transfer high-diversity structured knowledge from non-social recommendation models to social recommendation models. DivSR is a lightweight, model-agnostic framework that seamlessly integrates with existing social recommendation architectures. Experiments on three benchmark datasets demonstrate that DivSR significantly enhances diversity while maintaining competitive accuracy, achieving a superior accuracy-diversity trade-off. Our code and data are publicly available at: https://github.com/ll0ruc/DivSR.
We then introduce a novel hierarchical knowledge distillation strategy that incorporates the similarity matrix, feature representation, and response map-based distillation to guide the learning of the student Transformer network. We also enhance the model's ability to capture temporal dependencies by applying the temporal Fourier transform to establish temporal relationships between video frames. We adapt the network model to specific target objects during testing via a newly proposed test-time tuning strategy to achieve high performance and flexibility in target tracking. Recognizing the limitations of existing event-based tracking datasets, which are predominantly low-resolution, we propose EventVOT, the first large-scale high-resolution event-based tracking dataset. It comprises 1141 videos spanning diverse categories such as pedestrians, vehicles, UAVs, ping pong, etc. Extensive experiments on both low-resolution (FE240hz, VisEvent, FELT), and our newly proposed high-resolution EventVOT dataset fully validated the effectiveness of our proposed method. Both the benchmark dataset and source code have been released on https://github.com/Event-AHU/EventVOT_Benchmark
Audio-visual representation learning is crucial for advancing multimodal speech processing tasks, such as lipreading and audio-visual speech recognition. Recently, speech foundation models (SFMs) have shown remarkable generalization capabilities across various speech-related tasks. Building on this progress, we propose an audio-visual representation learning model that leverages cross-modal knowledge distillation from SFMs. In our method, SFMs serve as teachers, from which multi-layer hidden representations are extracted using clean audio inputs. We also introduce a multi-teacher ensemble method to distill the student, which receives audio-visual data as inputs. A novel representational knowledge distillation loss is employed to train the student during pretraining, which is also applied during finetuning to further enhance the performance on downstream tasks. Our experiments utilized both a self-supervised SFM, WavLM, and a supervised SFM, iFLYTEK-speech. The results demonstrated that our proposed method achieved superior or at least comparable performance to previous state-of-the-art baselines across automatic speech recognition, visual speech recognition, and audio-visual speech recognition tasks. Additionally, comprehensive ablation studies and the visualization of learned representations were conducted to evaluate the effectiveness of our proposed method.
This report details Lumen Labs' novel approach to processing Social Networking Service (SNS) data. We leverage knowledge distillation, specifically a simple distillation method inspired by DeepSeek-R1's CoT acquisition, combined with prompt hacking, to extract valuable training data from the Grok model. This data is then used to fine-tune a Phi-3-mini model, augmented with a mask-like mechanism specifically designed for handling the nuances of SNS data. Our method demonstrates state-of-the-art (SOTA) performance on several SNS data processing tasks, outperforming existing models like Grok, Phi-3, and GPT-4. We provide a comprehensive analysis of our approach, including mathematical formulations, engineering details, ablation studies, and comparative evaluations.
Federated Learning is vulnerable to adversarial manipulation, where malicious clients can inject poisoned updates to influence the global model's behavior. While existing defense mechanisms have made notable progress, they fail to protect against adversaries that aim to induce targeted backdoors under different learning and attack configurations. To address this limitation, we introduce DROP (Distillation-based Reduction Of Poisoning), a novel defense mechanism that combines clustering and activity-tracking techniques with extraction of benign behavior from clients via knowledge distillation to tackle stealthy adversaries that manipulate low data poisoning rates and diverse malicious client ratios within the federation. Through extensive experimentation, our approach demonstrates superior robustness compared to existing defenses across a wide range of learning configurations. Finally, we evaluate existing defenses and our method under the challenging setting of non-IID client data distribution and highlight the challenges of designing a resilient FL defense in this setting.
Multilingual neural machine translation (MNMT) aims at using one single model for multiple translation directions. Recent work applies non-autoregressive Transformers to improve the efficiency of MNMT, but requires expensive knowledge distillation (KD) processes. To this end, we propose an M-DAT approach to non-autoregressive multilingual machine translation. Our system leverages the recent advance of the directed acyclic Transformer (DAT), which does not require KD. We further propose a pivot back-translation (PivotBT) approach to improve the generalization to unseen translation directions. Experiments show that our M-DAT achieves state-of-the-art performance in non-autoregressive MNMT.
One of the most significant challenges in Music Emotion Recognition (MER) comes from the fact that emotion labels can be heterogeneous across datasets with regard to the emotion representation, including categorical (e.g., happy, sad) versus dimensional labels (e.g., valence-arousal). In this paper, we present a unified multitask learning framework that combines these two types of labels and is thus able to be trained on multiple datasets. This framework uses an effective input representation that combines musical features (i.e., key and chords) and MERT embeddings. Moreover, knowledge distillation is employed to transfer the knowledge of teacher models trained on individual datasets to a student model, enhancing its ability to generalize across multiple tasks. To validate our proposed framework, we conducted extensive experiments on a variety of datasets, including MTG-Jamendo, DEAM, PMEmo, and EmoMusic. According to our experimental results, the inclusion of musical features, multitask learning, and knowledge distillation significantly enhances performance. In particular, our model outperforms the state-of-the-art models, including the best-performing model from the MediaEval 2021 competition on the MTG-Jamendo dataset. Our work makes a significant contribution to MER by allowing the combination of categorical and dimensional emotion labels in one unified framework, thus enabling training across datasets.
Federated Reinforcement Learning (FedRL) improves sample efficiency while preserving privacy; however, most existing studies assume homogeneous agents, limiting its applicability in real-world scenarios. This paper investigates FedRL in black-box settings with heterogeneous agents, where each agent employs distinct policy networks and training configurations without disclosing their internal details. Knowledge Distillation (KD) is a promising method for facilitating knowledge sharing among heterogeneous models, but it faces challenges related to the scarcity of public datasets and limitations in knowledge representation when applied to FedRL. To address these challenges, we propose Federated Heterogeneous Policy Distillation (FedHPD), which solves the problem of heterogeneous FedRL by utilizing action probability distributions as a medium for knowledge sharing. We provide a theoretical analysis of FedHPD's convergence under standard assumptions. Extensive experiments corroborate that FedHPD shows significant improvements across various reinforcement learning benchmark tasks, further validating our theoretical findings. Moreover, additional experiments demonstrate that FedHPD operates effectively without the need for an elaborate selection of public datasets.
Federated Learning (FL) enables collaborative model training across multiple devices while preserving data privacy. However, it remains susceptible to backdoor attacks, where malicious participants can compromise the global model. Existing defence methods are limited by strict assumptions on data heterogeneity (Non-Independent and Identically Distributed data) and the proportion of malicious clients, reducing their practicality and effectiveness. To overcome these limitations, we propose Robust Knowledge Distillation (RKD), a novel defence mechanism that enhances model integrity without relying on restrictive assumptions. RKD integrates clustering and model selection techniques to identify and filter out malicious updates, forming a reliable ensemble of models. It then employs knowledge distillation to transfer the collective insights from this ensemble to a global model. Extensive evaluations demonstrate that RKD effectively mitigates backdoor threats while maintaining high model performance, outperforming current state-of-the-art defence methods across various scenarios.
Designing the physical encoder is crucial for accurate image reconstruction in computational imaging (CI) systems. Currently, these systems are designed via end-to-end (E2E) optimization, where the encoder is modeled as a neural network layer and is jointly optimized with the decoder. However, the performance of E2E optimization is significantly reduced by the physical constraints imposed on the encoder. Also, since the E2E learns the parameters of the encoder by backpropagating the reconstruction error, it does not promote optimal intermediate outputs and suffers from gradient vanishing. To address these limitations, we reinterpret the concept of knowledge distillation (KD) for designing a physically constrained CI system by transferring the knowledge of a pretrained, less-constrained CI system. Our approach involves three steps: (1) Given the original CI system (student), a teacher system is created by relaxing the constraints on the student's encoder. (2) The teacher is optimized to solve a less-constrained version of the student's problem. (3) The teacher guides the training of the student through two proposed knowledge transfer functions, targeting both the encoder and the decoder feature space. The proposed method can be employed to any imaging modality since the relaxation scheme and the loss functions can be adapted according to the physical acquisition and the employed decoder. This approach was validated on three representative CI modalities: magnetic resonance, single-pixel, and compressive spectral imaging. Simulations show that a teacher system with an encoder that has a structure similar to that of the student encoder provides effective guidance. Our approach achieves significantly improved reconstruction performance and encoder design, outperforming both E2E optimization and traditional non-data-driven encoder designs.
Spiking Neural Networks (SNNs) are emerging as a brain-inspired alternative to traditional Artificial Neural Networks (ANNs), prized for their potential energy efficiency on neuromorphic hardware. Despite this, SNNs often suffer from accuracy degradation compared to ANNs and face deployment challenges due to fixed inference timesteps, which require retraining for adjustments, limiting operational flexibility. To address these issues, our work considers the spatio-temporal property inherent in SNNs, and proposes a novel distillation framework for deep SNNs that optimizes performance across full-range timesteps without specific retraining, enhancing both efficacy and deployment adaptability. We provide both theoretical analysis and empirical validations to illustrate that training guarantees the convergence of all implicit models across full-range timesteps. Experimental results on CIFAR-10, CIFAR-100, CIFAR10-DVS, and ImageNet demonstrate state-of-the-art performance among distillation-based SNNs training methods. Our code is available at https://github.com/Intelli-Chip-Lab/snn\_temporal\_decoupling\_distillation.
Gait recognition is an important biometric technique over large distances. State-of-the-art gait recognition systems perform very well in controlled environments at close range. Recently, there has been an increased interest in gait recognition in the wild prompted by the collection of outdoor, more challenging datasets containing variations in terms of illumination, pitch angles, and distances. An important problem in these environments is that of occlusion, where the subject is partially blocked from camera view. While important, this problem has received little attention. Thus, we propose MimicGait, a model-agnostic approach for gait recognition in the presence of occlusions. We train the network using a multi-instance correlational distillation loss to capture both inter-sequence and intra-sequence correlations in the occluded gait patterns of a subject, utilizing an auxiliary Visibility Estimation Network to guide the training of the proposed mimic network. We demonstrate the effectiveness of our approach on challenging real-world datasets like GREW, Gait3D and BRIAR. We release the code in https://github.com/Ayush-00/mimicgait.