Searcharxiv⌕ Search

arXiv subjects

Tao Guo

Publications and source records attributed to Tao Guo.

At least 19 recordsLinked to original sources
arXiv2026

SLAI T-Rex: Full-Parameter Post-training of the DeepSeek-V4 Family on Ascend SuperPOD

Full-parameter post-training of trillion-parameter-scale MoE models introduces substantial system-level challenges for large-scale distributed training, including severe memory pressure, non-overlapped communication overhead, and inefficient kernel execution. While most large-scale LLM training systems are built around GPU-based clusters, this report presents an end-to-end optimization practice on the Ascend NPU SuperPOD. Using the DeepSeek-V4 model family as the target workload, we develop a hierarchical optimization framework spanning model-level parallelism, computation-communication orchestration, and low-level kernel execution. The resulting system achieves 34.22% Model FLOPs Utilization (MFU) with a 2.93x improvement over the open-source baseline recipe while maintaining training stability. Building on this optimized infrastructure, we further establish a CPT and SFT workflow for complex Operations Research (OR) tasks. We refer to the integrated framework as SLAI T-Rex. Using DeepSeek-V4-Flash, we develop OR-oriented CPT and SFT data pipelines that combine collected domain resources with solver-verified synthetic optimization documents. The resulting dataset contains 10K high-quality SFT samples spanning four task categories and three problem representations. The specialized model achieves the highest average zero-shot Pass@1 score among the evaluated models, reaching 71.81% and outperforming GPT-5.4-Mini and the base DeepSeek-V4-Flash model by 3.98 and 11.27 percentage points, respectively. Overall, this work demonstrates a full-stack pathway from efficient trillion-parameter model post-training on Ascend infra to domain-specialized Flash models for solver-grounded mathematical modeling, advancing frontier-model systems for complex reasoning.

cs.CL↗
arXiv2026

Adaptive Oscillatory-State Alignment for Time Series Forecasting

Long-term time series forecasting benefits from inductive biases that expose recurring temporal structure. Existing periodic forecasting methods typically model recurrence through predefined periods, global spectral components, or fixed learnable templates. However, real-world temporal dynamics are rarely rigidly periodic: around a nominal cycle, oscillatory behavior often exhibits \emph{non-rigid periodicity} (NRP), where cycle magnitude, cycle alignment, and local cycle duration vary over time. Under these conditions, fixed-template periodic modeling can become fundamentally mismatched to the underlying temporal states. We propose AOSNet, a Hilbert-guided forecasting framework that reformulates periodic forecasting from fixed template matching to adaptive oscillatory-state alignment. AOSNet extracts analytic-signal descriptors from both the observed sequence and a learnable global oscillatory prior, then adaptively aligns local states through a descriptor-conditioned gate that selectively preserves reliable observations while softly correcting mismatched regions. The learned prior serves not as a rigid repeated template but as a flexible oscillatory reference interpreted through local state dynamics. Experiments on eight public benchmarks and two cloud workload traces demonstrate leading or highly competitive accuracy with a compact model size and low inference latency, supporting repeated forecasting settings such as capacity planning and autoscaling. Controlled synthetic studies that isolate cycle-magnitude and cycle-alignment variation and combine them with cycle-duration changes show that the advantage of oscillatory-state alignment increases as NRP intensifies.

cs.LG↗
arXiv2026

SAM-Flow: Source-Anchored Masked Flow for Training-Free Image Editing

Training-free image editing has recently attracted increasing attention due to its ability to modify real images using powerful pre-trained diffusion and flow-matching models without additional training. However, existing inversion-based and differential-flow-based methods usually perform global latent transport, which inevitably propagates editing effects to non-target regions and leads to background leakage. To address this problem, we propose SAM-Flow, a source-anchored masked flow framework for localized training-free image editing. Instead of updating the whole latent representation, SAM-Flow first uses a scout image and token-grounded attention maps to localize the editable semantic regions. It then applies differential velocity updates only within these regions, while anchoring the remaining areas to the source-image latent trajectory. To further improve spatial stability and boundary naturalness, we introduce a time-varying source-anchored projection mechanism with dynamic soft masks, transition regions, and temporal mask accumulation. The proposed method is plug-and-play and can be integrated with mainstream flow-matching backbones such as Stable Diffusion 3 and FLUX without any fine-tuning. Extensive qualitative and quantitative experiments demonstrate that SAM-Flow achieves accurate semantic editing while significantly improving background preservation, providing a simple and general localized editing paradigm for training-free image editing. Code is available at: https://github.com/chwbob/Sam-Flow.

cs.CV↗
arXiv2026

BAHSD: Bridging the Long-tail Gap via Adaptive Distillation in Black-box Sequential Recommendation

Sequential recommendation systems are widely adopted but often deployed as black-box APIs, which has driven recent interest in model extraction to replicate their capabilities locally. However, the long-tail distribution induces severe signal heterogeneity: dense head sequences trigger the solidification of teacher preference, biasing extraction toward local patterns, while sparse tail sequences yield flat, noisy predictions. Existing one-size-fits-all extraction overlooks this disparity, resulting in noise overfitting and suboptimal knowledge transfer. We propose BAHSD, a black-box adaptive distillation framework that handles signal heterogeneity via a multi-scale consistency probing mechanism to implicitly quantify signal reliability. Based on this, an adaptive hierarchical objective is designed: dynamic-temperature KL divergence mitigates preference solidification for high-confidence signals, while ranking consistency and InfoNCE contrastive learning provide noise-robust enhancement for low-confidence signals. BAHSD consistently outperforms baselines, achieving up to 4.98\% gain over the teacher and 80\%+ improvement on tail users, offering a plug-and-play solution for high-fidelity black-box recommendation extraction.

cs.IR↗
arXiv2026

ChannelKAN: Multi-Scale Dual-Domain Channel Prediction via Hybrid CNN-KAN Architecture

Accurate channel state information (CSI) prediction is essential for improving the reliability and spectral efficiency of massive MIMO-OFDM systems in high-mobility scenarios. Existing deep learning methods struggle to jointly capture short-term local variations and long-range nonlinear dependencies in CSI sequences. To address this challenge, we propose ChannelKAN, a hybrid CNN-KAN channel prediction model with multi-scale frequency domain information enhancement. The key insight is that CNNs and Kolmogorov-Arnold Networks (KANs) are naturally complementary: CNNs extract intra-time-step local spatial-frequency correlations, while KANs with learnable Chebyshev polynomial activations fit inter-time-step nonlinear temporal evolution in a holistic manner. Specifically, a dual-domain expansion module first generates complementary frequency-domain and delay-domain CSI representations. A multi-scale frequency information enhancement module then retains dominant spectral components at multiple scales to strengthen key features and suppress noise. Next, a CNN-KAN feature extraction module captures local correlations via cascaded convolutions and models long-range dependencies via Chebyshev KAN layers. Finally, a dual-domain fusion module adaptively integrates features from both branches to produce the prediction. Experiments on 3GPP-compliant QuaDRiGa datasets demonstrate that ChannelKAN outperforms RNN, LSTM, GRU, CNN, and Transformer baselines in normalized mean square error (NMSE), spectral efficiency (SE), and bit error rate (BER) across various velocities and signal-to-noise ratios. Ablation studies further confirm the effectiveness of each proposed module.

eess.SP↗
arXiv2026

An Algebraic Method for Full-Rank Characterization in Binary Linear Coding

In this paper, we develop a characteristic set (CS)-based method for deriving full-rank equivalence conditions of symbolic matrices over the binary field. Such full-rank conditions are of fundamental importance for many linear coding problems in communication and information theory. Building on the developed CS-based method, we present an algorithm called Binary Characteristic Set for Full Rank (BCSFR), which efficiently derives the full-rank equivalence conditions as the zeros of a series of characteristic sets. In other words, the BCSFR algorithm can characterize all feasible linear coding schemes for certain linear coding problems (e.g., linear network coding and distributed storage coding), where full-rank constraints are imposed on several symbolic matrices to guarantee decodability or other properties of the codes. The derived equivalence conditions can be used to simplify the optimization of coding schemes, since the intractable full-rank constraints in the optimization problem are explicitly characterized by simple triangular-form equality constraints.

cs.IT↗
arXiv2026

RecBundle: A Next-Generation Geometric Paradigm for Explainable Recommender Systems

Recommender systems are inherently dynamic feedback loops where prolonged local interactions accumulate into macroscopic structural degradation such as information cocoons. Existing representation learning paradigms are universally constrained by the assumption of a single flat space, forcing topologically grounded user associations and semantically driven historical interactions to be fitted within the same vector space. This excessive coupling of heterogeneous information renders it impossible for researchers to mechanistically distinguish and identify the sources of systemic bias. To overcome this theoretical bottleneck, we introduce Fiber Bundle from modern differential geometry and propose a novel geometric analysis paradigm for recommender systems. This theory naturally decouples the system space into two hierarchical layers: the base manifold formed by user interaction networks, and the fibers attached to individual user nodes that carry their dynamic preferences. Building upon this, we construct RecBundle, a framework oriented toward next-generation recommender systems that formalizes user collaboration as geometric connection and parallel transport on the base manifold, while mapping content evolution to holonomy transformations on fibers. From this foundation, we identify future application directions encompassing quantitative mechanisms for information cocoons and evolutionary bias, geometric meta-theory for adaptive recommendation, and novel inference architectures integrating large language models (LLMs). Empirical analysis on real-world MovieLens and Amazon Beauty datasets validates the effectiveness of this geometric framework.

cs.IR↗
arXiv2026

MO-SAE:Multi-Objective Stacked Autoencoders Optimization for Edge Anomaly Detection

Stacked AutoEncoders (SAE) have been widely adopted in edge anomaly detection scenarios. However, the resource-intensive nature of SAE can pose significant challenges for edge devices, which are typically resource-constrained and must adapt rapidly to dynamic and changing conditions. Optimizing SAE to meet the heterogeneous demands of real-world deployment scenarios, including high performance under constrained storage, low power consumption, fast inference, and efficient model updates, remains a substantial challenge. To address this, we propose an integrated optimization framework that jointly considers these critical factors to achieve balanced and adaptive system-level optimization. Specifically, we formulate SAE optimization for edge anomaly detection as a multi-objective optimization problem and propose MO-SAE (Multi-Objective Stacked AutoEncoders). The multiple objectives are addressed by integrating model clipping, multi-branch exit design, and a matrix approximation technique. In addition, a multi-objective heuristic algorithm is employed to effectively balance the competing objectives in SAE optimization. Our results demonstrate that the proposed MO-SAE delivers substantial improvements over the original approach. On the x86 architecture, it reduces storage space and power consumption by at least 50%, improves runtime efficiency by no less than 28%, and achieves an 11.8% compression rate, all while maintaining application performance. Furthermore, MO-SAE runs efficiently on edge devices with ARM architecture. Experimental results show a 15% improvement in inference speed, facilitating efficient deployment in cloud-edge collaborative anomaly detection systems.

cs.NE↗
arXiv2026

Characterizations of Conditional Mutual Independence: Equivalence and Implication

Conditional independence, and more generally conditional mutual independence, are central notions in probability theory. In their general forms, they include functional dependence as a special case. In this paper, we tackle two fundamental problems related to conditional mutual independence. Let $K$ and $K'$ be two conditional mutual independncies (CMIs) defined on a finite set of discrete random variables. We have obtained a necessary and sufficient condition for i) $K$ is equivalent to $K'$; ii) $K$ implies $K'$. These characterizations are in terms of a canonical form introduced for conditional mutual independence.

math.PR↗
arXiv2026

MiMo-V2-Flash Technical Report

We present MiMo-V2-Flash, a Mixture-of-Experts (MoE) model with 309B total parameters and 15B active parameters, designed for fast, strong reasoning and agentic capabilities. MiMo-V2-Flash adopts a hybrid attention architecture that interleaves Sliding Window Attention (SWA) with global attention, with a 128-token sliding window under a 5:1 hybrid ratio. The model is pre-trained on 27 trillion tokens with Multi-Token Prediction (MTP), employing a native 32k context length and subsequently extended to 256k. To efficiently scale post-training compute, MiMo-V2-Flash introduces a novel Multi-Teacher On-Policy Distillation (MOPD) paradigm. In this framework, domain-specialized teachers (e.g., trained via large-scale reinforcement learning) provide dense and token-level reward, enabling the student model to perfectly master teacher expertise. MiMo-V2-Flash rivals top-tier open-weight models such as DeepSeek-V3.2 and Kimi-K2, despite using only 1/2 and 1/3 of their total parameters, respectively. During inference, by repurposing MTP as a draft model for speculative decoding, MiMo-V2-Flash achieves up to 3.6 acceptance length and 2.6x decoding speedup with three MTP layers. We open-source both the model weights and the three-layer MTP weights to foster open research and community collaboration.

cs.CL↗
arXiv2025

MiMo-Audio: Audio Language Models are Few-Shot Learners

Existing audio language models typically rely on task-specific fine-tuning to accomplish particular audio tasks. In contrast, humans are able to generalize to new audio tasks with only a few examples or simple instructions. GPT-3 has shown that scaling next-token prediction pretraining enables strong generalization capabilities in text, and we believe this paradigm is equally applicable to the audio domain. By scaling MiMo-Audio's pretraining data to over one hundred million of hours, we observe the emergence of few-shot learning capabilities across a diverse set of audio tasks. We develop a systematic evaluation of these capabilities and find that MiMo-Audio-7B-Base achieves SOTA performance on both speech intelligence and audio understanding benchmarks among open-source models. Beyond standard metrics, MiMo-Audio-7B-Base generalizes to tasks absent from its training data, such as voice conversion, style transfer, and speech editing. MiMo-Audio-7B-Base also demonstrates powerful speech continuation capabilities, capable of generating highly realistic talk shows, recitations, livestreaming and debates. At the post-training stage, we curate a diverse instruction-tuning corpus and introduce thinking mechanisms into both audio understanding and generation. MiMo-Audio-7B-Instruct achieves open-source SOTA on audio understanding benchmarks (MMSU, MMAU, MMAR, MMAU-Pro), spoken dialogue benchmarks (Big Bench Audio, MultiChallenge Audio) and instruct-TTS evaluations, approaching or surpassing closed-source models. Model checkpoints and full evaluation suite are available at https://github.com/XiaomiMiMo/MiMo-Audio.

cs.CL↗
arXiv2025

Applicability of Electrical Conductivity Ratio Method to Complicated Band Structure and the Carrier Scattering Mechanisms of SnSe

The electrical conductivity ratio (ECR) method can be used to analyze carrier scattering mechanism (CSM) without the need of magenetic transport measurements. In this work, the applicability of the ECR method in the analysis of complex energy band structures is discussed. Combined with the thermoelectric properties of SnSe, the feasibility using ECR method of ideal single band transport model to study the CSM of semiconductor materials with complicated band structure is studied. The results indicate that ECR method is not only applicable to idea band structure as reported before but also to the complicated band structure. The analysis results of the CSM of single crystal SnSe by ECR method using ideal single-band model agree with the carrier mobility temperature dependence in the nonphase transition temperature range. The CSM along three direction of single crystal SnSe are different because of its anisotropic crystal structure. The difference between dislocation scattering (DS) and charged impurity scattering (CIS) is that DS is always accompanied by polar optical phonon scattering (POP), such as the CSM of SnSe along b axis direction. The difference between CIS and DS can be more easily distinguished by the ECR method than the carrier mobility temperature dependent method. DS and POP might be one of the approaches to improve thermoelectric property because the scattering factor for DS or POP is larger than that of acoustic phonon scattering (APS) and alloy scattering (AS). For polycrystalline SnSe, the carrier scattering mechanisms varies with the crystal structure and the temperature. This indicates that the ECR method can better reflect the variation of carrier scattering mechanisms with temperature compared to the carrier mobility temperature dependence method.

cond-mat.mtrl-sci↗
arXiv2025

$\alpha$-leakage Interpretation of R\'{e}nyi Capacity

For $\tilde{f}(t) = \exp(\frac{\alpha-1}{\alpha}t)$, this paper shows that the Sibson mutual information is an $\alpha$-leakage averaged over the adversary's $\tilde{f}$-mean relative information gain (on the secret) at elementary event of channel output $Y$ as well as the joint occurrence of elementary channel input $X$ and output $Y$. This interpretation is used to derive a sufficient condition that achieves a $\delta$-approximation of $\epsilon$-upper bounded $\alpha$-leakage. A $Y$-elementary $\alpha$-leakage is proposed, extending the existing pointwise maximal leakage to the overall R\'{e}nyi order range $\alpha \in [0,\infty)$. Maximizing this $Y$-elementary leakage over all attributes $U$ of channel input $X$ gives the R\'{e}nyi divergence. Further, the R\'{e}nyi capacity is interpreted as the maximal $\tilde{f}$-mean information leakage over both the adversary's malicious inference decision and the channel input $X$ (represents the adversary's prior belief). This suggests an alternating max-max implementation of the existing generalized Blahut-Arimoto method.

cs.IT↗
arXiv2025

NS-Pep: De novo Peptide Design with Non-Standard Amino Acids

Peptide drugs incorporating non-standard amino acids (NSAAs) offer improved binding affinity and improved pharmacological properties. However, existing peptide design methods are limited to standard amino acids, leaving NSAA-aware design largely unexplored. We introduce NS-Pep, a unified framework for co-designing peptide sequences and structures with NSAAs. The main challenge is that NSAAs are extremely underrepresented-even the most frequent one, SEP, accounts for less than 0.4% of residues-resulting in a severe long-tailed distribution. To improve generalization to rare amino acids, we propose Residue Frequency-Guided Modification (RFGM), which mitigates over-penalization through frequency-aware logit calibration, supported by both theoretical and empirical analysis. Furthermore, we identify that insufficient side-chain modeling limits geometric representation of NSAAs. To address this, we introduce Progressive Side-chain Perception (PSP) for coarse-to-fine torsion and location prediction, and Interaction-Aware Weighting (IAW) to emphasize pocket-proximal residues. Moreover, NS-Pep generalizes naturally to the peptide folding task with NSAAs, addressing a major limitation of current tools. Experiments show that NS-Pep improves sequence recovery rate and binding affinity by 6.23% and 5.12%, respectively, and outperforms AlphaFold3 by 17.76% in peptide folding success rate.

q-bio.BM↗
arXiv2025

Channel, Trend and Periodic-Wise Representation Learning for Multivariate Long-term Time Series Forecasting

Downsampling-based methods for time series forecasting have attracted increasing attention due to their superiority in capturing sequence trends. However, this approaches mainly capture dependencies within subsequences but neglect inter-subsequence and inter-channel interactions, which limits forecasting accuracy. To address these limitations, we propose CTPNet, a novel framework that explicitly learns representations from three perspectives: i) inter-channel dependencies, captured by a temporal query-based multi-head attention mechanism; ii) intra-subsequence dependencies, modeled via a Transformer to characterize trend variations; and iii) inter-subsequence dependencies, extracted by reusing the encoder with residual connections to capture global periodic patterns. By jointly integrating these levels, proposed method provides a more holistic representation of temporal dynamics. Extensive experiments demonstrate the superiority of the proposed method.

cs.CE↗
arXiv2025

UniView: Enhancing Novel View Synthesis From A Single Image By Unifying Reference Features

The task of synthesizing novel views from a single image is highly ill-posed due to multiple explanations for unobserved areas. Most current methods tend to generate unseen regions from ambiguity priors and interpolation near input views, which often lead to severe distortions. To address this limitation, we propose a novel model dubbed as UniView, which can leverage reference images from a similar object to provide strong prior information during view synthesis. More specifically, we construct a retrieval and augmentation system and employ a multimodal large language model (MLLM) to assist in selecting reference images that meet our requirements. Additionally, a plug-and-play adapter module with multi-level isolation layers is introduced to dynamically generate reference features for the target views. Moreover, in order to preserve the details of an original input image, we design a decoupled triple attention mechanism, which can effectively align and integrate multi-branch features into the synthesis process. Extensive experiments have demonstrated that our UniView significantly improves novel view synthesis performance and outperforms state-of-the-art methods on the challenging datasets.

cs.CV↗
arXiv2025

Fully heavy tetraquark states with a diquark-antidiquark configuration

The study of the fully heavy tetraquark state helps to better understand the characteristics of compact exotic hadrons formed by quarks and gluons. In this work, we systematically investigate the mass spectra and decay channels of fully heavy tetraquark states using a diquark-antidiquark model based on the one-gluon exchange process. The tetraquark states are treated as compact diquark-antidiquark configurations, denoted as $[QQ][\bar{Q}\bar{Q}]$, where $Q = c$ or $b$. By properly accounting for the Pauli principle, basis wave functions incorporating flavor, color, and spin are constructed. The required model parameters are derived from fitting conventional hadron experimental values and lattice QCD calculations. We calculate the mass spectra and decays of the fully heavy tetraquarks in the lower $S$ and $P$ waves. Our results do not support interpreting $X(6600)$, $X(6900)$, and $X(7200)$ as the $1S$-wave fully charmed tetraquark states. Nevertheless, our analysis suggests that $X(6200)$ can be a good candidate for the $1S$-wave fully charmed tetraquark state with quantum numbers $2^{++}$. Furthermore, we also identify several narrow fully heavy tetraquark states that may be observable in future experiments.

hep-ph↗
arXiv2025

On the Evolution of Federated Post-Training Large Language Models: A Model Accessibility View

Federated Learning (FL) enables training models across decentralized data silos while preserving client data privacy. Recent research has explored efficient methods for post-training large language models (LLMs) within FL to address computational and communication challenges. While existing approaches often rely on access to LLMs' internal information, which is frequently restricted in real-world scenarios, an inference-only paradigm (black-box FedLLM) has emerged to address these limitations. This paper presents a comprehensive survey on federated tuning for LLMs. We propose a taxonomy categorizing existing studies along two axes: model access-based and parameter efficiency-based optimization. We classify FedLLM approaches into white-box, gray-box, and black-box techniques, highlighting representative methods within each category. We review emerging research treating LLMs as black-box inference APIs and discuss promising directions and open challenges for future research.

cs.LG↗