arXiv subjects

Bo Liu

Publications and source records attributed to Bo Liu.

At least 307 records · Page 17Linked to original sources

arXiv2024

Longhorn: State Space Models are Amortized Online Learners

Modern large language models are built on sequence modeling via next-token prediction. While the Transformer remains the dominant architecture for sequence modeling, its quadratic decoding complexity in sequence length poses a major limitation. State-space models (SSMs) present a competitive alternative, offering linear decoding efficiency while maintaining parallelism during training. However, most existing SSMs rely on linear recurrence designs that appear somewhat ad hoc. In this work, we explore SSM design through the lens of online learning, conceptualizing SSMs as meta-modules for specific online learning problems. This approach links SSM design to formulating precise online learning objectives, with state transition rules derived from solving these objectives. Based on this insight, we introduce a novel deep SSM architecture, Longhorn, whose update resembles the closed-form solution for solving the online associative recall problem. Our experimental results show that Longhorn outperforms state-of-the-art SSMs, including the Mamba model, on standard sequence modeling benchmarks, language modeling, and vision tasks. Specifically, Longhorn achieves a 1.8x improvement in sample efficiency compared to Mamba, and can extrapolate over contexts that are up to 16x longer during inference.

cs.LG↗

arXiv2024

Fine-Grained Gradient Restriction: A Simple Approach for Mitigating Catastrophic Forgetting

A fundamental challenge in continual learning is to balance the trade-off between learning new tasks and remembering the previously acquired knowledge. Gradient Episodic Memory (GEM) achieves this balance by utilizing a subset of past training samples to restrict the update direction of the model parameters. In this work, we start by analyzing an often overlooked hyper-parameter in GEM, the memory strength, which boosts the empirical performance by further constraining the update direction. We show that memory strength is effective mainly because it improves GEM's generalization ability and therefore leads to a more favorable trade-off. By this finding, we propose two approaches that more flexibly constrain the update direction. Our methods are able to achieve uniformly better Pareto Frontiers of remembering old and learning new knowledge than using memory strength. We further propose a computationally efficient method to approximately solve the optimization problem with more constraints.

cs.LG↗

arXiv2024

Propagation of Shear Horizontal Acoustic Waves in a Quartz Plate with a Fluid Layer for Liquid Sensor Application

We study the propagation of shear horizontal (SH) acoustic waves in a quartz elastic plate in contact with a viscous fluid layer of a finite thickness as an acoustic wave sensor for measuring fluid viscosity or density. The first order elastic plate theory by Mindlin and the theory of Newtonian fluids are used. An equation for determining the dispersion relations of the SH waves is obtained. Approximate dispersion relations for long waves are given analytically. Numerical results showing the effects of the fluid on SH wave characteristics are presented.

physics.app-ph↗

arXiv2024

Asynchronous Local-SGD Training for Language Modeling

Local stochastic gradient descent (Local-SGD), also referred to as federated averaging, is an approach to distributed optimization where each device performs more than one SGD update per communication. This work presents an empirical study of {\it asynchronous} Local-SGD for training language models; that is, each worker updates the global parameters as soon as it has finished its SGD steps. We conduct a comprehensive investigation by examining how worker hardware heterogeneity, model size, number of workers, and optimizer could impact the learning performance. We find that with naive implementations, asynchronous Local-SGD takes more iterations to converge than its synchronous counterpart despite updating the (global) model parameters more frequently. We identify momentum acceleration on the global parameters when worker gradients are stale as a key challenge. We propose a novel method that utilizes a delayed Nesterov momentum update and adjusts the workers' local training steps based on their computation speed. This approach, evaluated with models up to 150M parameters on the C4 dataset, matches the performance of synchronous Local-SGD in terms of perplexity per update step, and significantly surpasses it in terms of wall clock time.

cs.LG↗

arXiv2024

Diversity-grounded Channel Prototypical Learning for Out-of-Distribution Intent Detection

In the realm of task-oriented dialogue systems, a robust intent detection mechanism must effectively handle malformed utterances encountered in real-world scenarios. This study presents a novel fine-tuning framework for large language models (LLMs) aimed at enhancing in-distribution (ID) intent classification and out-of-distribution (OOD) intent detection, which utilizes semantic matching with prototypes derived from ID class names. By harnessing the highly distinguishable representations of LLMs, we construct semantic prototypes for each ID class using a diversity-grounded prompt tuning approach. We rigorously test our framework in a challenging OOD context, where ID and OOD classes are semantically close yet distinct, referred to as \emph{near} OOD detection. For a thorough assessment, we benchmark our method against the prevalent fine-tuning approaches. The experimental findings reveal that our method demonstrates superior performance in both few-shot ID intent classification and near-OOD intent detection tasks.

cs.CL↗

arXiv2024

Charged dilatonic black holes in dilaton-massive gravity

In this paper, we focus on massive Einstein-dilaton gravity including the coupling of dilaton scalar field to massive graviton terms, and then derive static and spherically symmetric solutions of charged dilatonic black holes in four dimensional spacetime. We find that the dilatonic black hole could possess different horizon structures for some suitably parameters. Then, we also investigate the thermodynamic properties of charged dilatonic black holes where $f(r)$ approaches $+\infty$ and $-\infty$, respectively.

gr-qc↗

arXiv2024

Magnetic ground states in kagome YCu$_3$(OH)$_6$[(Cl$_x$Br$_{1-x}$)$_{3-y}$(OH)$_{y}$]

Quantum spin liquids represent exotic states of spin systems characterized by long-range entanglement and emergent fractionalized quasiparticles. It is generally believed that disorder is hostile to quantum spin liquids. In our study, we investigated the magnetic properties of a kagome system, YCu$_3$(OH)$_6$[(Cl$_x$Br$_{1-x}$)$_{3-y}$(OH)$_{y}$]. Within this system, some of the hexagons exhibit alternate bonds along the Cu-O-Cu exchange paths, while others remain uniform. We found that a long-range antiferromagnetic order emerges when uniform hexagons dominate. Conversely, a possible quantum-spin-liquid state arises when the number of alternate-bond hexagons exceeds about 2/3. Therefore, the alternate-bond hexagons, typically considered as disorders, actually serve as the building blocks of the quantum spin liquid in this system. Notably, the low-temperature properties of the quantum spin liquid are directly associated with the height of the out-of-plane yttrium ions, which may be linked to changes in superexchange energies. Our results suggest that understanding the magnetic ground states in this system lies beyond the theoretical framework of the Heisenberg model constructed on the kagome lattice.

cond-mat.str-el↗

arXiv2024

When Fairness Meets Privacy: Exploring Privacy Threats in Fair Binary Classifiers via Membership Inference Attacks

Previous studies have developed fairness methods for biased models that exhibit discriminatory behaviors towards specific subgroups. While these models have shown promise in achieving fair predictions, recent research has identified their potential vulnerability to score-based membership inference attacks (MIAs). In these attacks, adversaries can infer whether a particular data sample was used during training by analyzing the model's prediction scores. However, our investigations reveal that these score-based MIAs are ineffective when targeting fairness-enhanced models in binary classifications. The attack models trained to launch the MIAs degrade into simplistic threshold models, resulting in lower attack performance. Meanwhile, we observe that fairness methods often lead to prediction performance degradation for the majority subgroups of the training data. This raises the barrier to successful attacks and widens the prediction gaps between member and non-member data. Building upon these insights, we propose an efficient MIA method against fairness-enhanced models based on fairness discrepancy results (FD-MIA). It leverages the difference in the predictions from both the original and fairness-enhanced models and exploits the observed prediction gaps as attack clues. We also explore potential strategies for mitigating privacy leakages. Extensive experiments validate our findings and demonstrate the efficacy of the proposed method.

cs.LG↗

arXiv2024

Memory-Efficient LLM Training with Online Subspace Descent

Recently, a wide range of memory-efficient LLM training algorithms have gained substantial popularity. These methods leverage the low-rank structure of gradients to project optimizer states into a subspace using projection matrix found by singular value decomposition (SVD). However, convergence of these algorithms is highly dependent on the update rules of their projection matrix. In this work, we provide the \emph{first} convergence guarantee for arbitrary update rules of projection matrix. This guarantee is generally applicable to optimizers that can be analyzed with Hamiltonian Descent, including most common ones, such as LION, Adam. Inspired by our theoretical understanding, we propose Online Subspace Descent, a new family of subspace descent optimizer without SVD. Instead of updating the projection matrix with eigenvectors, Online Subspace Descent updates the projection matrix with online PCA. Online Subspace Descent is flexible and introduces only minimum overhead to training. We show that for the task of pretraining LLaMA models ranging from 60M to 7B parameters on the C4 dataset, Online Subspace Descent achieves lower perplexity and better downstream tasks performance than state-of-the-art low-rank training methods across different settings and narrows the gap with full-rank baselines.

cs.LG↗

arXiv2024

TaSL: Continual Dialog State Tracking via Task Skill Localization and Consolidation

A practical dialogue system requires the capacity for ongoing skill acquisition and adaptability to new tasks while preserving prior knowledge. However, current methods for Continual Dialogue State Tracking (DST), a crucial function of dialogue systems, struggle with the catastrophic forgetting issue and knowledge transfer between tasks. We present TaSL, a novel framework for task skill localization and consolidation that enables effective knowledge transfer without relying on memory replay. TaSL uses a novel group-wise technique to pinpoint task-specific and task-shared areas. Additionally, a fine-grained skill consolidation strategy protects task-specific knowledge from being forgotten while updating shared knowledge for bi-directional knowledge transfer. As a result, TaSL strikes a balance between preserving previous knowledge and excelling at new tasks. Comprehensive experiments on various backbones highlight the significant performance improvements of TaSL over existing state-of-the-art methods. The source code is provided for reproducibility.

cs.CL↗

arXiv2024

Category-Prompt Refined Feature Learning for Long-Tailed Multi-Label Image Classification

Real-world data consistently exhibits a long-tailed distribution, often spanning multiple categories. This complexity underscores the challenge of content comprehension, particularly in scenarios requiring Long-Tailed Multi-Label image Classification (LTMLC). In such contexts, imbalanced data distribution and multi-object recognition pose significant hurdles. To address this issue, we propose a novel and effective approach for LTMLC, termed Category-Prompt Refined Feature Learning (CPRFL), utilizing semantic correlations between different categories and decoupling category-specific visual representations for each category. Specifically, CPRFL initializes category-prompts from the pretrained CLIP's embeddings and decouples category-specific visual representations through interaction with visual features, thereby facilitating the establishment of semantic correlations between the head and tail classes. To mitigate the visual-semantic domain bias, we design a progressive Dual-Path Back-Propagation mechanism to refine the prompts by progressively incorporating context-related visual information into prompts. Simultaneously, the refinement process facilitates the progressive purification of the category-specific visual representations under the guidance of the refined prompts. Furthermore, taking into account the negative-positive sample imbalance, we adopt the Asymmetric Loss as our optimization objective to suppress negative samples across all classes and potentially enhance the head-to-tail recognition performance. We validate the effectiveness of our method on two LTMLC benchmarks and extensive experiments demonstrate the superiority of our work over baselines. The code is available at https://github.com/jiexuanyan/CPRFL.

cs.CV↗

arXiv2024

DeepSeek-Prover-V1.5: Harnessing Proof Assistant Feedback for Reinforcement Learning and Monte-Carlo Tree Search

We introduce DeepSeek-Prover-V1.5, an open-source language model designed for theorem proving in Lean 4, which enhances DeepSeek-Prover-V1 by optimizing both training and inference processes. Pre-trained on DeepSeekMath-Base with specialization in formal mathematical languages, the model undergoes supervised fine-tuning using an enhanced formal theorem proving dataset derived from DeepSeek-Prover-V1. Further refinement is achieved through reinforcement learning from proof assistant feedback (RLPAF). Beyond the single-pass whole-proof generation approach of DeepSeek-Prover-V1, we propose RMaxTS, a variant of Monte-Carlo tree search that employs an intrinsic-reward-driven exploration strategy to generate diverse proof paths. DeepSeek-Prover-V1.5 demonstrates significant improvements over DeepSeek-Prover-V1, achieving new state-of-the-art results on the test set of the high school level miniF2F benchmark ($63.5\%$) and the undergraduate level ProofNet benchmark ($25.3\%$).

cs.CL↗

arXiv2024

3DSS-Mamba: 3D-Spectral-Spatial Mamba for Hyperspectral Image Classification

Hyperspectral image (HSI) classification constitutes the fundamental research in remote sensing fields. Convolutional Neural Networks (CNNs) and Transformers have demonstrated impressive capability in capturing spectral-spatial contextual dependencies. However, these architectures suffer from limited receptive fields and quadratic computational complexity, respectively. Fortunately, recent Mamba architectures built upon the State Space Model integrate the advantages of long-range sequence modeling and linear computational efficiency, exhibiting substantial potential in low-dimensional scenarios. Motivated by this, we propose a novel 3D-Spectral-Spatial Mamba (3DSS-Mamba) framework for HSI classification, allowing for global spectral-spatial relationship modeling with greater computational efficiency. Technically, a spectral-spatial token generation (SSTG) module is designed to convert the HSI cube into a set of 3D spectral-spatial tokens. To overcome the limitations of traditional Mamba, which is confined to modeling causal sequences and inadaptable to high-dimensional scenarios, a 3D-Spectral-Spatial Selective Scanning (3DSS) mechanism is introduced, which performs pixel-wise selective scanning on 3D hyperspectral tokens along the spectral and spatial dimensions. Five scanning routes are constructed to investigate the impact of dimension prioritization. The 3DSS scanning mechanism combined with conventional mapping operations forms the 3D-spectral-spatial mamba block (3DMB), enabling the extraction of global spectral-spatial semantic representations. Experimental results and analysis demonstrate that the proposed method outperforms the state-of-the-art methods on HSI classification benchmarks.

cs.CV↗

arXiv2024

Nonlinearity-induced dynamical self-organized twisted-bilayer lattices in Bose-Einstein condensates

Creating crystal bilayers twisted with respect to each other would lead to large periodic supercell structures, which can support a wide range of novel electron correlated phenomena, where the full understanding is still under debate. Here, we propose a new scheme to realize a nonlinearity-induced dynamical self-organized twisted-bilayer lattice in an atomic Bose-Einstein condensate (BEC). The key idea here is to utilize the nonlinear effect from the intrinsic atomic interactions to couple different layers and induce a dynamical self-organized supercell structure, dramatically distinct from the conventional wisdom to achieve the static twisted-bilayer lattices. To illustrate that, we study the dynamics of a two-component BEC and show that the nonlinear interaction effect naturally emerged in the Gross-Pitaevskii equation of interacting bosonic ultracold atoms can dynamically induce both periodic (commensurable) and aperiodic (incommensurable) moiré structures. One of the interesting moiré phenomena, i.e., the flat-band physics, is shown through investigating the dynamics of the wave packet of BEC. Our proposal can be implemented using available state-of-the-art experimental techniques and reveal a profound connection between the nonlinearity and twistronics in cold atom quantum simulators.

cond-mat.quant-gas↗

arXiv2024

Anomalous symmetry protected blockade of skin effect in one-dimensional non-Hermitian lattice systems

The non-Hermitian skin effect (NHSE), an anomalous localization behavior of the bulk states, is an inherently non-Hermitian phenomenon, which can not find a counterpart in Hermitian systems. However, the fragility of NHSE has been revealed recently, such as the boundary sensitivity, and it stimulates a lot of studies on discussing the fate of that. Here we present a theorem which shows that the combined spatial reflection symmetry can be considered as a criterion in one-dimensional non-Hermitian systems to determine whether the NHSE can exist or not. Distinct from previous studies, our proposed criterion only relies on analyzing the symmetry of the system, freeing out other requirements, such as the information of the energy spectrum. Furthermore, by taking the non-Hermitian Kitaev chain as an example, we verify our theorem through both a mathematical proof via the non-Bloch band theory and the exact diagonalization numerical studies. Our results reveal a profound connection between the symmetry and the fate of NHSE.

quant-ph↗

arXiv2024

Non-chiral non-Bloch invariants and topological phase diagram in non-unitary quantum dynamics without chiral symmetry

The non-Bloch topology leads to the emergence of various counter-intuitive phenomena in non-Hermitian systems under the open boundary condition (OBC), which can not find a counterpart in Hermitian systems. However, in the non-Hermitian system without chiral symmetry, being ubiquitous in nature, exploring its non-Bloch topology has so far eluded experimental effort. Here by introducing the concept of non-chiral non-Bloch invariants, we theoretically predict and experimentally identify the non-Bloch topological phase diagram of a one-dimensional (1D) non-Hermitian system without chiral symmetry in discrete-time non-unitary quantum walks of single photons. Interestingly, we find that such topological invariants not only can distinguish topologically distinct gapped phases, but also faithfully capture the corresponding gap closing in open-boundary spectrum at the phase boundary. Different topological regions are experimentally identified by measuring the featured discontinuities of the higher moments of the walker's displacement, which amazingly match excellently with our defined non-Bloch invariants. Our work provides a useful platform to study the interplay among topology, symmetries and the non-Hermiticity.

quant-ph↗

arXiv2024

UGG: Unified Generative Grasping

Dexterous grasping aims to produce diverse grasping postures with a high grasping success rate. Regression-based methods that directly predict grasping parameters given the object may achieve a high success rate but often lack diversity. Generation-based methods that generate grasping postures conditioned on the object can often produce diverse grasping, but they are insufficient for high grasping success due to lack of discriminative information. To mitigate, we introduce a unified diffusion-based dexterous grasp generation model, dubbed the name UGG, which operates within the object point cloud and hand parameter spaces. Our all-transformer architecture unifies the information from the object, the hand, and the contacts, introducing a novel representation of contact points for improved contact modeling. The flexibility and quality of our model enable the integration of a lightweight discriminator, benefiting from simulated discriminative data, which pushes for a high success rate while preserving high diversity. Beyond grasp generation, our model can also generate objects based on hand information, offering valuable insights into object design and studying how the generative model perceives objects. Our model achieves state-of-the-art dexterous grasping on the large-scale DexGraspNet dataset while facilitating human-centric object design, marking a significant advancement in dexterous grasping research. Our project page is https://jiaxin-lu.github.io/ugg/.

cs.CV↗

arXiv2024

Feature Re-Embedding: Towards Foundation Model-Level Performance in Computational Pathology

Multiple instance learning (MIL) is the most widely used framework in computational pathology, encompassing sub-typing, diagnosis, prognosis, and more. However, the existing MIL paradigm typically requires an offline instance feature extractor, such as a pre-trained ResNet or a foundation model. This approach lacks the capability for feature fine-tuning within the specific downstream tasks, limiting its adaptability and performance. To address this issue, we propose a Re-embedded Regional Transformer (R$^2$T) for re-embedding the instance features online, which captures fine-grained local features and establishes connections across different regions. Unlike existing works that focus on pre-training powerful feature extractor or designing sophisticated instance aggregator, R$^2$T is tailored to re-embed instance features online. It serves as a portable module that can seamlessly integrate into mainstream MIL models. Extensive experimental results on common computational pathology tasks validate that: 1) feature re-embedding improves the performance of MIL models based on ResNet-50 features to the level of foundation model features, and further enhances the performance of foundation model features; 2) the R$^2$T can introduce more significant performance improvements to various MIL models; 3) R$^2$T-MIL, as an R$^2$T-enhanced AB-MIL, outperforms other latest methods by a large margin.The code is available at: https://github.com/DearCaat/RRT-MIL.

cs.CV↗