arXiv subjects

Han Guo

Publications and source records attributed to Han Guo.

At least 19 recordsLinked to original sources

arXiv2026

LoRetta: A Foundation Model and Extensive Dataset for Global-Scale Remote Sensing Dense Image Matching

Dense image matching establishes pixel-wise correspondences and underpins broad applications in computer vision and photogrammetry. However, extending dense matching to global-scale remote sensing remains challenging because image pairs may differ in acquisition time, season, viewpoint, spatial resolution, and land-cover state. The resulting large geometric offsets, partial overlap, and intrinsically unmatchable regions make direct dense correspondence prediction unreliable and inefficient. We thus reformulate dense matching as localization-and-registration: first localizing the matchable overlap and affine geometry, then refining dense residuals within the aligned frame. Based on this formulation, we propose LoRetta, a foundation model coupling matchability-aware affine localization with guided dense registration. We also introduce LEVIR-GM, a global-scale multi-temporal optical matching benchmark with dataset-native matchability labels (103K aligned, 827K augmented pairs, six continents, five years, 0.5-1024 m resolution). We further establish a unified evaluation protocol for sparse, semi-dense, and dense matchers. On LEVIR-GM, LoRetta achieves an area under the curve (AUC) of 83.3%, outperforming the strongest baseline RoMa v2 by 1.6 points, with larger percentage of correct keypoints (PCK) gains of 6.5 and 8.2 points at 1 and 2 pixels, while reducing inference latency by 47.8%. Astronaut-to-satellite and unmanned aerial vehicle (UAV)-to-satellite geolocalization experiments further demonstrate its transferability as a reusable geometric aligner.

cs.CV↗

arXiv2026

FootprintNet: State-Transition-Guided Dynamic Footprint Learning for Multi-temporal Remote Sensing Change Detection

Despite substantial progress in remote sensing multi-temporal change detection (MTCD), most existing MTCD methods still represent the dynamic process at each spatial location over the entire observation period using a single change category associated with the final observation. This implicit single-change assumption limits their ability to characterize regions of recurrent change closely related to human activities. To address this limitation, we introduce Urban Building Dynamics Detection (UBDD), which identifies building-change dynamic footprints, i.e., the temporal intervals in which changes occur, from multi-temporal imagery and produces pixel-wise classification masks. For regions undergoing two or more changes, UBDD introduces an independent multi-change class for unified representation, thereby enabling unified modeling of single- and multi-change processes. Furthermore, we propose FootprintNet, which abstracts building-change processes as interactions between latent states and actions, and imposes state-action transition constraints to guide the learning of causally coherent change trajectories. It further exploits temporal change-boundary cues to enhance feature contrast across boundary sides, thereby improving the discrimination among different dynamic footprints and enabling accurate detection of dynamic footprints. Moreover, we introduce the Building Change Dynamics Score (BCDS) to address the inability of conventional metrics to reflect the temporal proximity between predicted footprints and labels. It evaluates predictions according to their preservation of change semantics and temporal offsets from the corresponding labels. Extensive experiments on TSCD, MUDS, and WUSU demonstrate that FootprintNet outperforms current state-of-the-art methods. The code is available at https://github.com/zmoka-zht/FootprintNet.

cs.CV↗

arXiv2026

BioMatrix: Towards a Comprehensive Biological Foundation Model Spanning the Modality Matrix of Sequences, Structures, and Language

We present BioMatrix, the first multimodal foundation model that natively integrates sequences, structures, and natural language for both molecules and proteins within a single decoder-only architecture. Existing biological foundation models pursue native multimodality and broad entity coverage separately: those that fuse multiple modalities under a shared objective remain confined to a single entity type, while those spanning multiple entity types either omit explicit structural modeling or rely on adapter-based designs in which the model cannot natively generate the very modalities it can read. BioMatrix closes this gap by mapping molecular sequences (supporting both SMILES and SELFIES notations), molecular structures, protein sequences, protein structures, and natural language into a shared discrete token space through a unified tokenization scheme, so that all modalities are consumed and produced uniformly under a single next-token prediction objective -- without external encoders, projection adapters, or modality-specific output heads. Built upon the Qwen3 language model (1.7B and 4B), BioMatrix is continually pretrained on 304.4 billion tokens spanning general and domain-specific text, sequence and structure views of molecules and proteins, and cross-modal corpora that interleave biomolecular entities with scientific text and link distinct entities through molecule-protein and protein-protein interaction data. After tuning on a comprehensive suite of downstream applications covering 80 tasks across 6 categories -- encompassing single-entity and multi-entity understanding and generation tasks across and within modalities -- BioMatrix achieves state-of-the-art or competitive performance on 77 out of 80 tasks, demonstrating that a single, natively multimodal generalist model can effectively match or surpass specialized approaches across a wide range of biological tasks.

cs.CL↗

arXiv2026

Hasse Diagrams for Attention: A Partial Order Framework for Designing Transformer Masks

During the training of large Transformer models, attention masks regulate the scope and direction of information flow across a sequence. Numerous mask variants exist, and operators such as FlexAttention already support arbitrary attention masks. Nevertheless, a systematic formal analysis of the information-flow structure induced by arbitrary masks has been missing. This paper develops a complete theoretical framework. We prove that, with sufficient depth, the information flow of a multi-layer Transformer converges to a Hasse diagram -- a directed acyclic graph representing a partial order. Building on this, we recast the design of parallel training tasks as the problem of finding a minimal common supergraph of Hasse diagrams, and we establish a criterion for the minimal common supergraph. This yields a constructive method to derive attention masks directly from a family of tasks. Applying the framework, we design two novel masks: a block-generation attention mask that ensures training-inference consistency (Block Two-Stream Attention), and a fully supervised bidirectional attention mask (Butterfly Attention). These results demonstrate the framework's capacity to discover new structures.

cs.LG↗

arXiv2026

CODA: Rewriting Transformer Blocks as GEMM-Epilogue Programs

Transformer training systems are built around dense linear algebra, yet a nontrivial fraction of end-to-end time is spent on surrounding memory-bound operators. Normalization, activations, residual updates, reductions, and related computations repeatedly move large intermediate tensors through global memory while performing little arithmetic, making data movement an increasingly important bottleneck in otherwise highly optimized training stacks. We introduce CODA, a GPU kernel abstraction that expresses these computations as GEMM-plus-epilogue programs. CODA is based on the observation that many Transformer operators exposed as separate framework kernels can be algebraically reparameterized to execute while a GEMM output tile remains on chip, before it is written to memory. The abstraction fixes the GEMM mainloop and exposes a small set of composable epilogue primitives for scaling, reductions, pairwise transformations, and accumulation. This constrained interface preserves the performance structure of expert-written GEMMs while remaining expressive enough to cover nearly all non-attention computation in the forward and backward pass of a standard Transformer block. Across representative Transformer workloads, both human- and LLM-authored CODA kernels achieve high performance, suggesting that GEMM-plus-epilogue programming offers a practical path toward combining framework-level productivity with hardware-level efficiency.

cs.LG↗

arXiv2026

Fast KV Compaction via Attention Matching

Scaling language models to long contexts is often bottlenecked by the size of the key-value (KV) cache. In deployed settings, long contexts are typically managed through compaction in token space via summarization. However, summarization can be highly lossy, substantially harming downstream performance. Recent work on Cartridges has shown that it is possible to train highly compact KV caches in latent space that closely match full-context performance, but at the cost of slow and expensive end-to-end optimization. This work describes an approach for fast context compaction in latent space through Attention Matching, which constructs compact keys and values to reproduce attention outputs and preserve attention mass at a per-KV-head level. We show that this formulation naturally decomposes into simple subproblems, some of which admit efficient closed-form solutions. Within this framework, we develop a family of methods that significantly push the Pareto frontier of compaction time versus quality, achieving up to 50x compaction in seconds on some datasets with little quality loss.

cs.LG↗

arXiv2025

TaCo: Capturing Spatio-Temporal Semantic Consistency in Remote Sensing Change Detection

Remote sensing change detection (RSCD) aims to identify surface changes across bi-temporal satellite images. Most previous methods rely solely on mask supervision, which effectively guides spatial localization but provides limited constraints on the temporal semantic transitions. Consequently, they often produce spatially coherent predictions while still suffering from unresolved semantic inconsistencies. To address this limitation, we propose TaCo, a spatio-temporal semantic consistent network, which enriches the existing mask-supervised framework with a spatio-temporal semantic joint constraint. TaCo conceptualizes change as a semantic transition between bi-temporal states, in which one temporal feature representation can be derived from the other via dedicated transition features. To realize this, we introduce a Text-guided Transition Generator that integrates textual semantics with bi-temporal visual features to construct the cross-temporal transition features. In addition, we propose a spatio-temporal semantic joint constraint consisting of bi-temporal reconstruct constraints and a transition constraint: the former enforces alignment between reconstructed and original features, while the latter enhances discrimination for changes. This design can yield substantial performance gains without introducing any additional computational overhead during inference. Extensive experiments on six public datasets, spanning both binary and semantic change detection tasks, demonstrate that TaCo consistently achieves SOTA performance.

cs.CV↗

arXiv2025

FoBa: A Foreground-Background co-Guided Method and New Benchmark for Remote Sensing Semantic Change Detection

Despite the remarkable progress achieved in remote sensing semantic change detection (SCD), two major challenges remain. At the data level, existing SCD datasets suffer from limited change categories, insufficient change types, and a lack of fine-grained class definitions, making them inadequate to fully support practical applications. At the methodological level, most current approaches underutilize change information, typically treating it as a post-processing step to enhance spatial consistency, which constrains further improvements in model performance. To address these issues, we construct a new benchmark for remote sensing SCD, LevirSCD. Focused on the Beijing area, the dataset covers 16 change categories and 210 specific change types, with more fine-grained class definitions (e.g., roads are divided into unpaved and paved roads). Furthermore, we propose a foreground-background co-guided SCD (FoBa) method, which leverages foregrounds that focus on regions of interest and backgrounds enriched with contextual information to guide the model collaboratively, thereby alleviating semantic ambiguity while enhancing its ability to detect subtle changes. Considering the requirements of bi-temporal interaction and spatial consistency in SCD, we introduce a Gated Interaction Fusion (GIF) module along with a simple consistency loss to further enhance the model's detection performance. Extensive experiments on three datasets (SECOND, JL1, and the proposed LevirSCD) demonstrate that FoBa achieves competitive results compared to current SOTA methods, with improvements of 1.48%, 3.61%, and 2.81% in the SeK metric, respectively. Our code and dataset are available at https://github.com/zmoka-zht/FoBa.

cs.CV↗

arXiv2025

Self-Adapting Language Models

Large language models (LLMs) are powerful but static; they lack mechanisms to adapt their weights in response to new tasks, knowledge, or examples. We introduce Self-Adapting LLMs (SEAL), a framework that enables LLMs to self-adapt by generating their own finetuning data and update directives. Given a new input, the model produces a self-edit-a generation that may restructure the information in different ways, specify optimization hyperparameters, or invoke tools for data augmentation and gradient-based updates. Through supervised finetuning (SFT), these self-edits result in persistent weight updates, enabling lasting adaptation. To train the model to produce effective self-edits, we use a reinforcement learning loop with the downstream performance of the updated model as the reward signal. Unlike prior approaches that rely on separate adaptation modules or auxiliary networks, SEAL directly uses the model's own generation to control its adaptation process. Experiments on knowledge incorporation and few-shot generalization show that SEAL is a promising step toward language models capable of self-directed adaptation. Our website and code is available at https://jyopari.github.io/posts/seal.

cs.LG↗

arXiv2025

Log-Linear Attention

The attention mechanism in Transformers is an important primitive for accurate and scalable sequence modeling. Its quadratic-compute and linear-memory complexity however remain significant bottlenecks. Linear attention and state-space models enable linear-time, constant-memory sequence modeling and can moreover be trained efficiently through matmul-rich parallelization across sequence length. However, at their core these models are still RNNs, and thus their use of a fixed-size hidden state to model the context is a fundamental limitation. This paper develops log-linear attention, an attention mechanism that balances linear attention's efficiency and the expressiveness of softmax attention. Log-linear attention replaces the fixed-size hidden state with a logarithmically growing set of hidden states. We show that with a particular growth function, log-linear attention admits a similarly matmul-rich parallel form whose compute cost is log-linear in sequence length. Log-linear attention is a general framework and can be applied on top of existing linear attention variants. As case studies, we instantiate log-linear variants of two recent architectures -- Mamba-2 and Gated DeltaNet -- and find they perform well compared to their linear-time variants.

cs.LG↗

arXiv2025

Reinforcement Learning Based Symbolic Regression for Load Modeling

With the increasing penetration of renewable energy sources, growing demand variability, and evolving grid control strategies, accurate and efficient load modeling has become a critical yet challenging task. Traditional methods, such as fixed-form parametric models and data-driven approaches, often struggle to balance accuracy, computational efficiency, and interpretability. This paper introduces a novel symbolic regression algorithm based on the Actor-Critic reinforcement learning framework, specifically tailored for dynamic load modeling. The algorithm employs a trainable expression tree with controlled depth and a predefined set of operators to generate compact and interpretable mathematical expressions. The Actor network probabilistically selects operators for the symbolic expression, while the Critic evaluates the resulting expression tree through a loss function. To further enhance performance, a candidate pool mechanism is implemented to store high-performing expressions, which are subsequently fine-tuned using gradient descent. By focusing on simplicity and precision, the proposed method significantly reduces computational complexity while preserving interpretability. Experimental results validate its superior performance compared to existing benchmarks, which offers a robust and scalable solution for dynamic load modeling and system analysis in modern power systems.

eess.SY↗

arXiv2025

DALNet: A Denoising Diffusion Probabilistic Model for High-Fidelity Day-Ahead Load Forecasting

Accurate probabilistic load forecasting is crucial for maintaining the safety and stability of power systems. However, the mainstream approach, multi-step prediction, is hindered by cumulative errors and forecasting lags, which limits its effectiveness in probabilistic day-ahead load forecasting (PDALF). To overcome these challenges, we introduce DALNet, a novel denoising diffusion model designed to generate load curves rather than relying on direct prediction. By shifting the focus to curve generation, DALNet captures the complex distribution of actual load time-series data under specific conditions with greater fidelity. To further enhance DALNet, we propose the temporal multi-scale attention block (TMSAB), a mechanism designed to integrate both positional and temporal information for improved forecasting precision. Furthermore, we utilize kernel density estimation (KDE) to reconstruct the distribution of generated load curves and employ Kullback-Leibler (KL) divergence to compare them with the actual data distribution. Experimental results demonstrate that DALNet excels in load forecasting accuracy and offers a novel perspective for other predictive tasks within power systems.

eess.SY↗

arXiv2025

On the Duality between Gradient Transformations and Adapters

We study memory-efficient optimization of neural networks (in particular language models) with linear gradient transformations, where the gradients are linearly mapped to a lower dimensional space than the full parameter space, thus saving memory required for gradient accumulation and optimizer state persistence. The model parameters are updated by first performing an optimization step in the lower dimensional space and then going back into the original parameter space via the linear map's transpose. We show that optimizing the model in this transformed space is equivalent to reparameterizing the original model through a linear adapter that additively modifies the model parameters, and then only optimizing the adapter's parameters. When the transformation is Kronecker-factored, this establishes an equivalence between GaLore and one-sided LoRA. We show that this duality between gradient transformations and adapter-based reparameterizations unifies existing approaches to memory-efficient training and suggests new techniques for improving training efficiency and memory use.

cs.LG↗

arXiv2024

Pushing the Limits of Large Language Model Quantization via the Linearity Theorem

Quantizing large language models has become a standard way to reduce their memory and computational costs. Typically, existing methods focus on breaking down the problem into individual layer-wise sub-problems, and minimizing per-layer error, measured via various metrics. Yet, this approach currently lacks theoretical justification and the metrics employed may be sub-optimal. In this paper, we present a "linearity theorem" establishing a direct relationship between the layer-wise $\ell_2$ reconstruction error and the model perplexity increase due to quantization. This insight enables two novel applications: (1) a simple data-free LLM quantization method using Hadamard rotations and MSE-optimal grids, dubbed HIGGS, which outperforms all prior data-free approaches such as the extremely popular NF4 quantized format, and (2) an optimal solution to the problem of finding non-uniform per-layer quantization levels which match a given compression constraint in the medium-bitwidth regime, obtained by reduction to dynamic programming. On the practical side, we demonstrate improved accuracy-compression trade-offs on Llama-3.1 and 3.2-family models, as well as on Qwen-family models. Further, we show that our method can be efficiently supported in terms of GPU kernels at various batch sizes, advancing both data-free and non-uniform quantization for LLMs.

cs.LG↗

arXiv2024

The Surprising Effectiveness of Test-Time Training for Few-Shot Learning

Language models (LMs) have shown impressive performance on tasks within their training distribution, but often struggle with structurally novel tasks even when given a small number of in-context task examples. We investigate the effectiveness of test-time training (TTT) -- temporarily updating model parameters during inference using a loss derived from input data -- as a mechanism for improving LMs' reasoning and few-shot learning capabilities. On the Abstraction and Reasoning Corpus (ARC), performing TTT with in-context examples yields up to $6\times$ higher accuracy compared to fine-tuned baselines -- reaching $53.0\%$ on the public validation set with an 8B-parameter LM and $61.9\%$ when ensembled with program-synthesis methods, matching average human performance. On BIG-Bench Hard (BBH), TTT on in-context examples surpasses standard few-shot prompting in the $10$-shot setting by $7.3$ percentage points ($50.5\%$ to $57.8\%$). Our findings highlight the limitations of in-context learning for novel tasks and demonstrate the potential of test-time training to enhance language model adaptability.

cs.AI↗

arXiv2024

Convolution tensor decomposition for efficient high-resolution solutions to the Allen-Cahn equation

This paper presents a convolution tensor decomposition based model reduction method for solving the Allen-Cahn equation. The Allen-Cahn equation is usually used to characterize phase separation or the motion of anti-phase boundaries in materials. Its solution is time-consuming when high-resolution meshes and large time scale integration are involved. To resolve these issues, the convolution tensor decomposition method is developed, in conjunction with a stabilized semi-implicit scheme for time integration. The development enables a powerful computational framework for high-resolution solutions of Allen-Cahn problems, and allows the use of relatively large time increments for time integration without violating the discrete energy law. To further improve the efficiency and robustness of the method, an adaptive algorithm is also proposed. Numerical examples have confirmed the efficiency of the method in both 2D and 3D problems. Orders-of-magnitude speedups were obtained with the method for high-resolution problems, compared to the finite element method. The proposed computational framework opens numerous opportunities for simulating complex microstructure formation in materials on large-volume high-resolution meshes at a deeply reduced computational cost.

math.NA↗

arXiv2024

Training-Free Activation Sparsity in Large Language Models

Activation sparsity can enable practical inference speedups in large language models (LLMs) by reducing the compute and memory-movement required for matrix multiplications during the forward pass. However, existing methods face limitations that inhibit widespread adoption. Some approaches are tailored towards older models with ReLU-based sparsity, while others require extensive continued pre-training on up to hundreds of billions of tokens. This paper describes TEAL, a simple training-free method that applies magnitude-based activation sparsity to hidden states throughout the entire model. TEAL achieves 40-50% model-wide sparsity with minimal performance degradation across Llama-2, Llama-3, and Mistral families, with sizes varying from 7B to 70B. We improve existing sparse kernels and demonstrate wall-clock decoding speed-ups of up to 1.53$\times$ and 1.8$\times$ at 40% and 50% model-wide sparsity. TEAL is compatible with weight quantization, enabling further efficiency gains.

cs.CL↗

arXiv2024

Fast Matrix Multiplications for Lookup Table-Quantized LLMs

The deployment of large language models (LLMs) is often constrained by memory bandwidth, where the primary bottleneck is the cost of transferring model parameters from the GPU's global memory to its registers. When coupled with custom kernels that fuse the dequantization and matmul operations, weight-only quantization can thus enable faster inference by reducing the amount of memory movement. However, developing high-performance kernels for weight-quantized LLMs presents substantial challenges, especially when the weights are compressed to non-evenly-divisible bit widths (e.g., 3 bits) with non-uniform, lookup table (LUT) quantization. This paper describes FLUTE, a flexible lookup table engine for LUT-quantized LLMs, which uses offline restructuring of the quantized weight matrix to minimize bit manipulations associated with unpacking, and vectorization and duplication of the lookup table to mitigate shared memory bandwidth constraints. At batch sizes < 32 and quantization group size of 128 (typical in LLM inference), the FLUTE kernel can be 2-4x faster than existing GEMM kernels. As an application of FLUTE, we explore a simple extension to lookup table-based NormalFloat quantization and apply it to quantize LLaMA3 to various configurations, obtaining competitive quantization performance against strong baselines while obtaining an end-to-end throughput increase of 1.5 to 2 times.

cs.LG↗