Deep neural networks often underperform on tabular data due to sensitivity to irrelevant features and a spectral bias toward smooth, low-frequency functions, limiting their ability to capture sharp, high-frequency signals in low-label regimes. While self-supervised learning (SSL) holds promise in such settings, it remains challenging in tabular domains due to the limited availability of effective data augmentations. We introduce TANDEM (Tree-And-Neural Dual Encoder Model), a hybrid autoencoder that trains a neural encoder alongside an oblivious soft decision tree (OSDT) encoder, both guided by dedicated stochastic gating networks for sample-specific feature selection. The encoders share a decoder and are coupled via alignment losses, encouraging complementary yet consistent representations. The training-only use of the tree operates as model-based augmentation, nudging representations toward tabular-relevant features while preserving a lean inference path (only the neural encoder is deployed). Spectral analysis highlights distinct yet complementary inductive biases across encoders, and experiments on classification and regression benchmarks in low-label settings show consistent gains over strong deep, tree-based, and SSL baselines.

Gilgo Beach serial killer Rex Heuermann's ex-wife reacts to his sentencing as monster who killed eight women is transferred to new prison to begin life behind bars Boy, three, 'attacked by at least one crocodile' after being'thrown into zoo pit by man with learning difficulties who broke away from carers' - as suspect'not fit for interview' is bailed Jelly Roll stops concert to respond to wife Bunnie XO's bombshell podcast on their divorce Hegseth puts NATO on notice as he launches review of US troops in Europe and blasts allies for'shameful' behavior I was mortified when my husband always said no to sex. Then I realised the mistake I was making. This is the change that's completely transformed marital love-making in middle age: ALICE SNAPE Mom thought popular'natural' health supplement was safer than Xanax. She took it... then never woke up. Don't make the same mistake JD Vance turns on Israeli allies who are criticizing Trump's Iran deal: 'Wake up and smell reality' The other women left devastated by Jelly Roll's divorce... why his daughter is now'disgusted'... and Bunnie XO's one red-line demand before she would agree to the split Joe Biden mumbles to himself and requires stage direction as he aimlessly wanders off at Obama's library debut Tourists run for their lives as gunfire erupts in New York's Times Square as terrified parents drag children to safety Heartbroken family of college girls who drowned dispute account of their final moments before they were swept out to sea as they mourn'responsible and kind' students Oscar-winning director's daughter and her husband's deaths'medically related' as cops give grim update after couple were found in SUV on California highway Furious woke woman storms out of restaurant because customers were singing National Anthem ...and vows never to return A bold new experiment to streamline how Americans buy new cars... and auto dealerships are already scared Secret White House blacklist leaked by insider: 'Worst' influencers named and shamed... as foul-mouthed backstabbing erupts Watch horrifying drone video that follows woman's plunge to death after bungee team threw her from bridge without rope Bill Clinton's VERY cozy moment with Michelle while Hillary looks the other way... and the best UNSEEN moments from Obama public library opening Farce of Obama's $850m'monstrosity': As clucking liberal elite cheer Barack's grand opening, outraged Chicago locals tell HARRIET ALEXANDER awkward truth about library Humiliating new joke about Trump that's the talk of Washington... as White House moles tell me there's more to this story than meets the eye: MARK HALPERIN Humanoid robot is spotted BEGGING on a street in China - claiming it has'no money to recharge' READ MORE: China unveils the world's first self-driving TOILET While many people worry that robots are coming to take their jobs, one unlucky bot seems to have fallen on hard times.

Text-to-image diffusion models rely on text embeddings from a pre-trained text encoder, but these embeddings remain fixed across all diffusion timesteps, limiting their adaptability to the generative process. We propose Diffusion Adaptive Text Embedding (DATE), which dynamically updates text embeddings at each diffusion timestep based on intermediate perturbed data. We formulate an optimization problem and derive an update rule that refines the text embeddings at each sampling step to improve alignment and preference between the mean predicted image and the text. This allows DATE to dynamically adapts the text conditions to the reverse-diffused images throughout diffusion sampling without requiring additional model training. Through theoretical analysis and empirical results, we show that DATE maintains the generative capability of the model while providing superior text-image alignment over fixed text embeddings across various tasks, including multi-concept generation and text-guided image editing.

Large vision-language models (LVLMs) excel at multimodal understanding but suffer from high computational costs due to redundant vision tokens. Existing pruning methods typically rely on single-layer attention scores to rank and prune redundant visual tokens to solve this inefficiency. However, as the interaction between tokens and layers is complicated, this raises a basic question: Is such a simple single-layer criterion sufficient to identify redundancy? To answer this question, we rethink the emergence of redundant visual tokens from a fundamental perspective: information flow, which models the interaction between tokens and layers by capturing how information moves between tokens across layers. We find (1) the CLS token acts as an information relay, which can simplify the complicated flow analysis; (2) the redundancy emerges progressively and dynamically via layer-wise attention concentration; and (3) relying solely on attention scores from single layers can lead to contradictory redundancy identification. Based on this, we propose FlowCut, an information-flow-aware pruning framework, mitigating the insufficiency of the current criterion for identifying redundant tokens and better aligning with the model's inherent behaviors. Extensive experiments show that FlowCut achieves superior results, outperforming SoTA by 1.6% on LLaVA-1.57B with 88.9% token reduction, and by 4.3% on LLaVA-NeXT-7B with 94.4% reduction, delivering 3.2 speed-up in the prefilling stage.

Conditional generative modeling aims to learn a conditional data distribution from samples containing data-condition pairs. For this, diffusion and flow-based methods have attained compelling results. These methods use a learned (flow) model to transport an initial standard Gaussian noise that ignores the condition to the conditional data distribution. The model is hence required to learn both mass transport and conditional injection. To ease the demand on the model, we propose Condition-Aware Reparameterization for Flow Matching (CAR-Flow) - a lightweight, learned shift that conditions the source, the target, or both distributions. By relocating these distributions, CAR-Flow shortens the probability path the model must learn, leading to faster training in practice. On low-dimensional synthetic data, we visualize and quantify the effects of CAR-Flow. On higher-dimensional natural image data (ImageNet-256), equipping SiT-XL/2 with CAR-Flow reduces FID from 2.07 to 1.68, while introducing less than 0.6% additional parameters.

Stochastic Gradient Descent (SGD) and its Ruppert-Polyak averaged variant (ASGD) lie at the heart of modern large-scale learning, yet their theoretical properties in high-dimensional settings are rarely understood. In this paper, we provide rigorous statistical guarantees for constant learning-rate SGD and ASGD in high-dimensional regimes. Our key innovation is to transfer powerful tools from high-dimensional time series to online learning. Specifically, by viewing SGD as a nonlinear autoregressive process and adapting existing coupling techniques, we prove the geometric-moment contraction of high-dimensional SGD for constant learning rates, thereby establishing asymptotic stationarity of the iterates. Building on this, we derive the q-th moment convergence of SGD and ASGD for any q 2 in general ℓs-norms, and, in particular, the ℓ -norm that is frequently adopted in high-dimensional sparse or structured models. Furthermore, we provide sharp high-probability concentration analysis which entails the probabilistic bound of high-dimensional ASGD. Beyond closing a critical gap in SGD theory, our proposed framework offers a novel toolkit for analyzing a broad class of high-dimensional learning algorithms.

Inspired by the success of DeepSeek-R1, we explore the potential of rule-based reinforcement learning (RL) in MLLM post-training for perception policy learning. While promising, our initial experiments reveal that incorporating a thinking process through RL does not consistently lead to performance gains across all visual perception tasks. This leads us to delve into the essential role of RL in the context of visual perception. In this work, we return to the fundamentals and explore the effects of RL on different perception tasks. We observe that the perceptual perplexity is a major factor in determining the effectiveness of RL. We also observe that reward design plays a crucial role in further approaching the upper limit of model perception. To leverage these findings, we propose Perception-R1, a scalable RL framework using GRPO during MLLM post-training.

The reasoning capabilities of large language models (LLMs) have advanced rapidly, particularly following the release of DeepSeek-R1, which has inspired a surge of research into data quality and reinforcement learning (RL) algorithms. Despite the pivotal role diversity plays in RL, its influence on LLM reasoning remains largely underexplored. To bridge this gap, this work presents a systematic investigation into the impact of diversity in RL-based training for LLM reasoning, and proposes a novel diversity-aware policy optimization method. Across evaluations on 12 LLMs, we observe a strong positive correlation between the solution diversity and Potential@k (a novel metric quantifying an LLM's reasoning potential) in high-performing models. This finding motivates our method to explicitly promote diversity during RL training. Specifically, we design a token-level diversity and reformulate it into a practical objective, then we selectively apply it to positive samples. Integrated into the R1-zero training framework, our method achieves a 3.5% average improvement across four mathematical reasoning benchmarks, while generating more diverse and robust solutions.

Federated Learning (FL) confronts a significant challenge known as data heterogeneity, which impairs model performance and convergence. Existing methods have made notable progress in addressing this issue. However, improving performance in certain heterogeneity scenarios remains an overlooked question: How robust are these methods to deploy under diverse heterogeneity scenarios? To answer this, we conduct comprehensive evaluations across varied heterogeneity scenarios, showing that most existing methods exhibit limited robustness. Meanwhile, insights from these experiments highlight that sharing statistical information can mitigate heterogeneity by enabling clients to update with a global perspective. Motivated by this, we propose FedGPS (Federated Goal-Path Synergy), a novel framework that seamlessly integrates statistical distribution and gradient information from others. Specifically, FedGPS statically modifies each client's learning objective to implicitly model the global data distribution using surrogate information, while dynamically adjusting local update directions with gradient information from other clients at each round. Extensive experiments show that FedGPS outperforms state-of-the-art methods across diverse heterogeneity scenarios, validating its effectiveness and robustness.

Audio captioning systems face a fundamental challenge: teacher-forcing training creates exposure bias that leads to caption degeneration during inference. While contrastive methods have been proposed as solutions, they typically fail to capture the crucial temporal relationships between acoustic and linguistic modalities. We address this limitation by introducing the unbiased sliced Wasserstein RBF (USWRBF) kernel with rotary positional embedding, specifically designed to preserve temporal information across modalities. Our approach offers a practical advantage: the kernel enables efficient stochastic gradient optimization, making it computationally feasible for real-world applications. Building on this foundation, we develop a complete audio captioning framework that integrates stochastic decoding to further mitigate caption degeneration. Extensive experiments on AudioCaps and Clotho datasets demonstrate that our method significantly improves caption quality, lexical diversity, and text-to-audio retrieval accuracy. Furthermore, we demonstrate the generalizability of our USW-RBF kernel by applying it to audio reasoning tasks, where it enhances the reasoning capabilities of large audio language models on the CompA-R in terms of correctness and quality. Our kernel also improves the reasoning accuracy of the MMAU-test-mini benchmarks by 4%. These results establish our approach as a powerful and generalizable solution for cross-modal alignment challenges in audio-language tasks.