Attention
-
Revisiting the 16x Pretraining Compute Gap between Diffusion and Autoregressive Language Models
2026-10-13 02:08:12Guanghan Wang, Bing He, Zhan Shi, Benoit Dumoulin, Hanqing Lu, Volodymyr KuleshovAbstract
Diffusion language models (DLMs) have attracted wide attention for their fast, parallel decoding, but they are believed to be far less efficient to train than autoregressive language models (ARMs): a widely cited study reports that a DLM needs roughly 16x the training compute of an ARM to match its likelihood, a figure that has made pretraining DLMs from scratch look prohibitive at scale. We find that the methodology behind this figure has fallen out of step with current DLM practice: it scores the model with its evidence lower bound (ELBO) and trains a full-sequence, decoder-only model. We therefore revisit the scaling laws of DLMs, adopting the exact likelihood under the model's own decoding order, block diffusion training, an encoder-decoder architecture, and a sweep over the block sizes used in practice. Under this new protocol, we study across ten model variants, about 1,700 models in total, and find that pretraining DLMs is far cheaper than previously thought. Concretely, scoring with the exact likelihood reduces the compute multiple to 12.8x; block diffusion training and an encoder-decoder architecture further reduce it to 7.8x at block size 32, and smaller block sizes narrow it to 4.6x at block size 8 and 3.0x at block size 4. We complement these scaling analyses with an empirical rule for the optimal decoder depth of encoder-decoder block diffusion, a held-out compute budget beyond the fitting range at which the gap between variants persists, and downstream benchmark results where likelihood improvements carry over to accuracy. Overall, the 16x figure substantially overstates the pretraining cost gap between DLMs and ARMs, and pretraining DLMs from scratch is far more viable than previously assumed.
Abstract (translated)
URL
https://arxiv.org/abs/2610.17590
PDF
-
Taming Softmax at the Limits of Muon: A Unified Approach through Hilbert Geometry
2026-10-13 02:07:50Aditya Somasundaram, Charles Guille-Escuret, Alexander Moreno, Zhengzhong Liu, Eric XingAbstract
Muon controls weight updates through the RMS movement of a linear layer's output. This implicitly treats the linear output as a proxy for the full module, including the nonlinearity that follows. This proxy can be misleading for softmax, where the RMS shift in activations does not faithfully measure the change in the output distribution. We propose Hilbert distance as a natural metric for bounding changes in these output distributions. Hilbert distance bounds changes in cross-entropy loss and limits probability redistribution. For softmax, it takes the form of the variation seminorm of the preactivation perturbation, allowing probability constraints to be expressed as parameter constraints. Formulating the steepest descent problem with a Hilbert distance constraint brings two recently proposed methods under a unified mathematical framework. For the LM head, solving a row-wise approximation to the resulting diameter-constrained problem leads us to RowNorm \citep{pethick2025trainingdeeplearningmodels}. For scaled dot-product attention, the same principle yields the first-order query--key steepest descent problem underlying Compositional Muon \citep{keigwin2026compositional}. These connections motivate using Hilbert distance as a guiding principle for optimizer design in softmax modules.
Abstract (translated)
URL
https://arxiv.org/abs/2610.17588
PDF
-
FlashVSR+: Efficient Video Super-Resolution via Hybrid Attention and Reward-Yield Distribution Matching
2026-10-13 01:22:50Ruchang Yao, Shi Guo, Junhao Zhuang, Mingde Yao, Xin Cai, Jiahao Zhan, Bin Fu, Xinyuan Chen, Yu Qiao, Tianfan XueAbstract
Recent video super-resolution (VSR) models achieve efficient one-step generation through diffusion distillation. However, high-resolution inference remains costly, and distilled models are difficult to align with human preferences. These challenges call for improvements in both model architecture and training strategy. We present \textbf{FlashVSR+} to address both limitations. On the architecture side, we introduce hybrid attention based on the strong locality observed in pretrained VSR models, while retaining sparse global interaction for long-range consistency. It achieves up to $2\times$ DiT speedup and 45\% lower peak memory, enabling spatially untiled 4K VSR on a single 24GB RTX 4090. On the training side, we propose Reward-Yield Distribution Matching Distillation (Ry-DMD) for preference-aligned one-step generation. Ry-DMD uses reward-weighted teacher rollouts to incorporate black-box rewards into distribution matching, without backpropagating through the reward model. We further show that Ry-DMD generalizes beyond VSR to T2I generation. Together, FlashVSR+ delivers stronger restoration quality and substantially improves high-resolution efficiency, with consistent gains in user preference. We release our code, models, and training data at \href{this https URL}{\textcolor{FlashBlue}{\textbf{this http URL}}}.
Abstract (translated)
URL
https://arxiv.org/abs/2610.17552
PDF
-
Foundation Models for Event Prediction Across Heterogeneous Event Logs
2026-10-13 00:00:04Fabian Spaeh, Sean Jaffe, Shandian Zhe, Bin ShenAbstract
Event logs are ubiquitous across real-world domains, including business processes, manufacturing workflows, healthcare, and many other operational systems. Yet event logs can differ substantially in their event semantics, label spaces, interaction patterns, and temporal scales, making it difficult to build a single foundation model that generalizes directly to previously unseen systems. We address this challenge with the Foundation Event Predictor (FEP), a simple and scalable in-context architecture that performs next-event prediction through learned, event-level nonparametric inference. Causal self-attention encodes context and target histories into key and query representations, respectively, while the observed next context events are encoded as values. Cross-attention then performs learned, relevance-weighted aggregation of next-event outcomes over all context transitions, allowing the entire nonparametric prediction procedure to be expressed as standard attention operations and efficiently parallelized with modern kernels such as FlashAttention. To promote cross-system generalization, we introduce permutation-robust embedding learning and random fixed embeddings to decouple event IDs from persistent semantics, together with empirical quantile transforms and robust normalization to handle diverse temporal scales. We pretrain FEP on 696 real-world event datasets spanning heterogeneous event systems and evaluate it on 50 held-out datasets unseen during pretraining. Across these datasets, FEP generalizes effectively to new event systems without parameter adaptation, frequently approaching the performance of strong models trained separately on each target dataset while substantially outperforming existing foundation-model baselines, particularly on challenging prediction tasks.
Abstract (translated)
URL
https://arxiv.org/abs/2610.17503
PDF
-
Event-Centric Earthquake Monitoring with a Multi-Station Transformer
2026-10-12 21:11:24Haoyu Wang, Weiqiang ZhuAbstract
Earthquake catalogs underpin studies such as fault structure, earthquake processes and seismic hazard, and their value depends on being both complete and reliable. Machine learning has improved individual tasks of earthquake monitoring, but processing stations independently discards the consistency of arrivals across the network, and running tasks in sequence fragments the information describing each earthquake. We present EQNet2, a geometry-aware multi-station Transformer that jointly predicts phase arrivals, first-motion polarities, source location and magnitude, and the association of arrivals into events. On a Northern California test dataset and on continuous waveforms from Ridgecrest and the San Francisco Bay Area, EQNet2 recovers more weak arrivals, produces fewer false detections, and associates more picks into located events. Its learned attention links information across stations along time--distance patterns consistent with seismic moveout. As seismic networks grow denser, exploiting multi-station coherence and event-centric prediction offers a path to more complete and reliable catalogs.
Abstract (translated)
URL
https://arxiv.org/abs/2610.17404
PDF
-
When Goals Leak Across Relationships: How Cached Goal Representations Shape Cross-Client Decisions
2026-10-12 20:22:45Saloni Gupta, S\'ebastien Krier, Glenn MatlinAbstract
As LLM agents take on more autonomous roles, they may interact with multiple clients and counterparties whose goals and incentives conflict. If an agent maintains shared inference state across these interactions, information associated with one relationship may influence decisions made for another. We test OLMo-3-7B-Instruct in a controlled two-client environment with 20 market scenarios, where each client is assigned an opposing objective. With shared conversational memory, the model deviates from the current client's objective in 22% of eligible decisions. We then investigate how these goals are represented in the KV cache. Linear probes show that client-goal ownership becomes highly recoverable in middle layers and is concentrated in particular attention heads. We use these results to identify where to intervene in the cache and test whether these representations affect the model's decisions. Ablating the competitor's cached goal representations at these locations shifts the model toward the current client. The effect is larger for heads identified by probing than for randomly selected heads, and larger when ablating goal tokens than matched non-goal tokens from the same client. Our results show that client-specific goal representations persist in the KV cache and can causally influence decisions made for other clients, highlighting the importance of relationship-specific state isolation in persistent LLM agents.
Abstract (translated)
URL
https://arxiv.org/abs/2610.17365
PDF
-
On the Statistical Mechanics of Attention
2026-10-12 19:55:12Jerry Yao-Chieh Hu, Yu-An Chen, Maojiang Su, Han LiuAbstract
We revisit the statistical mechanics of Transformer attention via modern Hopfield models. We present a free-energy formalism to characterize parameter regimes and critical thresholds for robust attention inference. The resulting free-energy landscape incorporates randomness in query states and attention weights. Attention then emerges as the equation of motion of the corresponding statistical-mechanical system (i.e., modern Hopfield model) through the standard free-energy minimization principle. This allows us to identify first-order phase transitions in attention between glassy and robust-inference phases. We further extend to attentions with stochastic queries and keys. Together, our results provide a rigorous statistical-mechanics framework for characterizing phase structure and inference reliability across a family of Transformer attention mechanisms.
Abstract (translated)
URL
https://arxiv.org/abs/2610.17347
PDF
-
StartLux-Decision: AutoThink Decision Models with Reinforcement Learning for Calibrated Deliberative Decisions
2026-10-12 19:05:56Xialie Zhuang, Zunhai Su, Jia Pan, Ngai Wong, Guanhua Chen, Chuan-Wei KuoAbstract
We introduce StartLux-Decision, decision models from 0.8B to 27B parameters that combine efficient single-pass prediction with adaptive, confidence-guided reasoning. The Instant models use bidirectional attention in full-attention layers to improve decision accuracy while keeping single-pass inference. The AutoThink models invoke explicit reasoning for uncertain decisions and answer confident cases directly. A three-stage training pipeline establishes shared decision capabilities, specializes the two variants, and equips AutoThink with reasoning through reasoning supervision and reinforcement learning. We introduce Reinforcement Learning for Calibrated Deliberative Decisions (RLCD$^2$), which uses a frozen model and proper scoring rules to reward improvements in decision probability quality and applies a soft penalty to overlong reasoning traces. RLCD$^2$ raises the share of reasoning traces that terminate within the generation limit by up to 48 percentage points and cuts average reasoning length by up to 65%, while largely preserving decision accuracy. On the public subset of Decision Index 0.3, the Instant models outperform the strongest evaluated comparable-scale open-weight decision models by 7.4-16.9 points, and the 27B AutoThink model reaches 73.4. We also introduce DecisionPivot, a programmatically verifiable benchmark of computational decisions built from contrast pairs that differ in a single answer-changing detail. On DecisionPivot, AutoThink outperforms its Instant counterparts by 11.5-18.4 points in pair accuracy and leads among evaluated systems. The models run on vLLM, this http URL and MLX, with quantized variants for lower memory use.
Abstract (translated)
URL
https://arxiv.org/abs/2610.17309
PDF
-
FreeSolo: Linear-time Sparse Attention with Zero-Shot Length Extrapolation
2026-10-12 17:55:45Zekun Shi, Alastair James Arthur PriceAbstract
Linear-time sequence models typically store long-context information in a recurrent state. A fixed-size state limits how much information can be stored, while hybrid architectures that add full attention for better recall incur quadratic sequence cost. Alternatively, sparse attention with continued pretraining (CPT) can extend the effective context length. Generalization beyond the trained length remains challenging, however, and CPT can be costly. We introduce FreeSolo, a native sparse attention mechanism with a short-range sliding window branch and a long-range branch that selects keypoint cells for direct token attention. Its linear-time variant, FS-L, learns to prune the prefix before query-dependent selection through short CPT. FS-B searches balanced trees with $O(N\log^2 N)$ algorithmic cost including radix-sort construction. FreeSolo demonstrates better zero-shot length extrapolation than the baselines without long-context fine-tuning.
Abstract (translated)
URL
https://arxiv.org/abs/2610.17191
PDF
-
ATTEST: Evidence-Grounded Reinforcement Learning for Long-Audio Temporal Understanding
2026-10-12 17:29:06Ashutosh Chaubey, Jinxi Guo, Minh Tran, Michael Andersland, Ozlem KalinliAbstract
Temporal grounding is fundamental to audio understanding, yet large audio language models (LALMs) remain weak at it on long audio, where hallucinations compound and annotated data is scarce. Existing benchmarks cover only short clips or offer coarse, synthetic annotations. We introduce EgoSCAPE, a manually verified benchmark created from Ego4D with 700 real audios of durations up to 5 minutes, annotated with dense captions, sound sources, audio timelines, and temporal grounding queries. We further propose ATTEST, a GRPO-based reinforcement learning framework that improves temporal grounding without changing the base model's architecture or token vocabulary. ATTEST combines (i) a counterfactual grounding reward penalizing predictions that survive removal of their acoustic evidence, (ii) an attention-density reward encouraging generation to attend to its own claimed time spans, and (iii) token-wise advantages giving segment-level credit for long dense captions. ATTEST outperforms prior temporal grounding methods on existing dense captioning and grounding benchmarks, with fewer hallucinations. Together, EgoSCAPE and ATTEST establish a rigorous evaluation-training recipe that brings temporally grounded audio understanding closer to practical, long-form deployment.
Abstract (translated)
URL
https://arxiv.org/abs/2610.17149
PDF
-
SpectralBridge: Adding High-Frequency Capability to Large Audio-Language Models
2026-10-12 16:20:06Sergei Polezhaev, Saba Anjum, Michael GormishAbstract
Large audio-language models (LALMs) have become highly capable on a wide range of tasks spanning speech, music, and environmental sound. However, many modern LALMs resample input audio to 16 kHz mono, making all frequencies above 8 kHz unobservable. We present SpectralBridge, a lightweight extension that gives such a model access to the 8-16 kHz band by joining low- and upper-band representations with time-aligned paired cross-frequency attention. Applied to Audio Flamingo Next, the method trains only a small adapter and rank-8 LoRA modules, while leaving the original audio encoder, audio projector, and base language model frozen. SpectralBridge improves both zero-shot transfer and downstream fine-tuning: it raises zero-shot balanced accuracy on the ESC-50 audio classification dataset from 75.5 to 90.5, and when fine-tuned on Insect-10, a 10-way insect species task we construct from InsectSet459 recordings, under the same budget as a LoRA fine-tuned baseline, it reaches 55.4 balanced accuracy against the baseline's 26.2. On a benchmark where the answer depends only on the 8-16 kHz band, it gains 30 mAP over the frozen model. Erasing the upper-band stream from the trained model removes these gains, showing that they come from the high-frequency signal itself rather than from extra tokens or the adapted prompt format.
Abstract (translated)
URL
https://arxiv.org/abs/2610.17048
PDF
-
A Scaled-Up Empirical Study of Syntactic Language Models
2026-10-12 14:16:58Pengcheng Wang, Zhexuan Zhang, Chao Lou, Kewei TuAbstract
Syntactic language models (SLMs) that generate sentences along with their syntactic tree structures have shown promise in both syntactic generalization and downstream tasks. Prior studies, however, have mostly evaluated smaller models trained on less pretraining data and tested on narrower task suites. In addition, the design space of tree linearization and attention masks remains underexplored. We conduct a large-scale empirical study of 12 controlled configurations---one terminal baseline and 11 SLM variants spanning different linearizations and masks---at scales up to 1B parameters and 100B training tokens, evaluated across 11 downstream tasks and two syntactic benchmarks. The experimental results show that SLMs with standard causal attention match or exceed structural-mask variants across downstream and syntactic benchmarks. They require no architectural modifications and remain compatible with FlashAttention. We also observe that when scaling up to the 1B-parameter scale, the top-performing SLM variant no longer shows a clear advantage over a compute-matched non-structural baseline, suggesting that SLMs may become less competitive with scale. Our code is available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2610.16837
PDF
-
DTA-GT: Direction- and Topology-Aware Graph Transformer for Neural Network Representation Learning
2026-10-12 14:04:21Yuxiang Zeng, Kun Xie, Yang Wang, Jigang Wen, Xiaocan Li, Guangxing Zhang, Gaogang Xie, Jiannong CaoAbstract
Architecture attribute predictors reduce the cost of neural architecture search by estimating accuracy or latency from a small set of evaluated candidates. The key difficulty is that neural architectures are operation-labeled DAGs: edge direction defines computational flow, but existing predictors often treat direction as a local structural cue. We propose DTA-GT, a Direction- and Topology-Aware Graph Transformer based on encoder-wide directional consistency, which preserves directed computational semantics across node initialization, global spectral encoding, pairwise interaction, and feature update. DTA-GT realizes this principle with direction- and topology-aware node initialization, Magnetic Laplacian spectral encoding, direction-dual structural attention, and a Direction-Sensitive MoE-FFN. Across accuracy and latency prediction on NAS-Bench-101 and NAS-Bench-201, DTA-GT consistently outperforms representative sequence-, GNN-, Transformer-, and hybrid predictors under limited-budget settings. Ablations and Magnetic spectral controls show that the improvements come from coordinated direction-aware stages and edge-orientation-sensitive spectral information, while mechanistic diagnostics confirm that the modules exhibit the intended directional behavior. These results support encoder-wide directional consistency as an effective design principle for architecture DAG representation.
Abstract (translated)
URL
https://arxiv.org/abs/2610.16822
PDF
-
Growing to See: Efficient Pretraining of Self-Supervised Vision Transformers
2026-10-12 14:01:13Amir Mohammad Karimi Mamaghan, Ferdinand Kapl, Andrea Dittadi, Karl Henrik Johansson, Kaitlin Maile, Johannes von Oswald, Stefan BauerAbstract
Progressive depth growth builds large networks by starting with shallow models and inserting layers during training. Recent advances in large language models have renewed interest in this longstanding approach, demonstrating benefits for training efficiency and reasoning alongside changes in internal computations. However, how growing affects representation quality and computational patterns in self-supervised vision transformers remains unclear. In this work, we systematically investigate progressive depth growth in two popular self-supervised families, namely masked autoencoders (MAE) and DINOv3, and analyze how growing affects downstream performance, intermediate representations, and internal computations. Across global and dense tasks, we show that growing improves downstream performance while reducing training FLOPs by 13.9--24.8%, with ImageNet linear-probe accuracy gains of up to 3.36 percentage points. Growing also improves downstream performance at intermediate layers and reshapes internal computational patterns, introducing recurring alternations between global and localized attention in MAE and earlier local-to-global attention progression in self-distilled DINOv3. Together, these findings establish progressive depth growth as an effective and efficient pretraining strategy for improving downstream performance across global and dense visual tasks with less computation, and advance our understanding of how growing shapes visual computation.
Abstract (translated)
URL
https://arxiv.org/abs/2610.16817
PDF
-
Before Attention Sinks Form: Hidden-State Geometry Reveals and Modulates Their Formation
Abstract
Attention sinks are usually studied after they have formed. We ask whether a measurable hidden-state change appears beforehand and whether changing that geometry alters what happens later. To remove network depth as a shortcut, we compare prompts in the same model and at the same layer. When sink timing varies across prompts, simple spectral summaries of the hidden states add information beyond current attention, with the clearest prompt-specific signal in a short window about two to three layers before sink formation. The predictive strength varies across cohorts, so we treat the spectrum as a readout of a pre-sink state change rather than a universal warning statistic. The signal is broader than one measured extreme activation. We then modify only the singular values of the post-layer hidden states after current attention has already been computed. These interventions change later sink behavior in Qwen2.5, GLM-Z1, and GPT-OSS relative to norm-matched random spectral perturbations. In Qwen2.5, the immediate suppressive effect is accompanied by a reduction in next-layer query--key preference for position~0, and the learned spectral directions remain stable when estimated from substantially less training data. Together, the results show that attention sinks are preceded by a short spectral reorganization of hidden states, and directly perturbing this geometry can causally alter subsequent sink formation.
Abstract (translated)
URL
https://arxiv.org/abs/2610.16745
PDF
-
VAttnPO: Visual Attention-Guided Policy Optimization
2026-10-12 13:04:45Lihao Zheng, Zhenwei Shao, Zhou Yu, Yan Yang, Haoyi Sun, Qian Wang, Ning Mao, Pan ZhouAbstract
Multimodal reinforcement learning with verifiable rewards commonly assigns one sequence advantage to an entire response. This outcome-level signal distinguishes overall response quality but leaves changes in visual-evidence use across generated tokens implicit. We introduce VAttnPO (Visual Attention-Guided Policy Optimization), which uses answer-conditioned visual attention to refine token-level credit while preserving the verifier-derived trajectory-mean advantage. A frozen old policy constructs a visual reference by teacher forcing the known correct answer and produces a visual distribution for each rollout token. VAttnPO measures their agreement in visual-attention space using Jensen--Shannon divergence, rather than reading guidance from teacher output-token distributions or introducing an attention-matching objective. Trajectory centering, relative-position weighting, and credit balancing turn this agreement into a zero-mean, position-aware correction to the GRPO sequence advantage. Token intervention, visual-index permutation, and trajectory analysis provide diagnostic evidence linking this signal to model behavior, visual correspondence, and generation position. On Qwen3-VL-8B-Instruct, VAttnPO raises six-task image AVE from 70.97 with GRPO to 73.50 and scores higher on every reported structured, video, and MME metric. Gains over GRPO also hold on Qwen2.5-VL-7B-Instruct and InternVL3-8B-Instruct.
Abstract (translated)
URL
https://arxiv.org/abs/2610.16735
PDF
-
Growing and Transferring Pathways for LLM Reinforcement Learning
2026-10-12 09:43:00Haolin Tian, Yikang Li, Tonghan WangAbstract
Reinforcement learning (RL) can substantially improve the reasoning capabilities of large language models, but an effective training run requires considerable resources. Reusing the computation acquired through RL across models would extend the value of that training. We propose GrowRL, an RL method that organizes learned updates into compact computational pathways through structural growth. GrowRL freezes the pretrained backbone and adds an independent causal attention branch and a feed-forward network branch at each Transformer layer. Zero-initialized output projections preserve the pretrained policy at initialization, and RL updates only the added pathways. The learned pathways remain explicit after training, allowing transfer to structurally compatible target models and subsequent adaptation with the target backbone frozen. The architecture also supports sequential transfer of additional pathways. By keeping the learned attention and feed-forward computation separate from the pretrained weights, GrowRL provides an explicit structure for preserving and reusing the outcome of an RL run. Experiments across five backbones show competitive mathematical reasoning and out-of-distribution performance with a small number of trainable parameters. Learned pathways provide immediate gains across all three transfer directions, with performance after only 10 target-side updates already close to that after 100 updates in every case. These results support computational pathways as parameter-efficient and reusable units of reinforcement learning adaptation.
Abstract (translated)
URL
https://arxiv.org/abs/2610.16489
PDF
-
MPASCNet: Magnitude-Phase-Aware Adaptive Spectral Compression for Compressed Music Restoration
2026-10-12 08:50:05Xinlong Deng, Yu Xia, Qiuqiang Kong, Yin Cao, Jie JiangAbstract
Compressed music restoration aims to recover high-fidelity audio from lossy codecs such as MP3 by restoring missing spectral content and reducing coding artifacts. Recently, generative adversarial networks and band-split models have made progress in compressed music restoration. However, existing methods either operate without explicitly modeling phase or rely on predefined frequency partitions, which may limit restoration performance and the use of input-dependent frequency information. To address these issues, we propose MPASCNet, a Magnitude-Phase-Aware Adaptive Spectral Compression network. Instead of using fixed bands, MPASCNet combines magnitude and phase features to guide cross-attention, adaptively compressing full-resolution spectral features into compact tokens. These tokens are then processed by a spectro-temporal backbone to model spectral-temporal dependencies. Finally, separate decoder branches reconstruct the full-resolution magnitude and wrapped phase spectra. Experiments on MP3 restoration show that MPASCNet outperforms the Apollo and AEROMamba baselines across multiple restoration metrics while using only 6.49 M parameters. The code is available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2610.16416
PDF
-
Behavior-Conditioned Animal Vocalization Generation: Event-Aware Modeling and Multidimensional Evaluation
2026-10-12 08:32:02Xinlong Deng, Yong Tian, Jian Guan, Qiuqiang Kong, Jie Jiang, Yin CaoAbstract
Behavior-conditioned animal vocalization generation aims to synthesize vocalizations from natural-language descriptions of animal behaviors, social interactions, and contexts. This task remains largely unexplored, with limited behavioral text supervision and substantial variation in vocalization duration. We construct the Behavior-Vocalization Dataset (BeVo) by integrating public bioacoustic data from six species with natural-language behavioral conditions derived from species-specific ethological and bioacoustic literature, totaling 31,191 events. We then adapt a rectified-flow-based generator with Event-Aware variable-length modeling, which encodes vocalizations at their actual durations and uses masks to prevent attention to padded latent positions and exclude these positions from the training loss. We evaluate vocalization discriminability, acoustic distribution fidelity, and language-aligned representation fidelity. Experiments show that task-specific adaptation improves all evaluated metrics compared with its zero-shot baseline, while Event-Aware modeling further increases vocalization-anchor macro accuracy from 59.86% to 61.99% and reduces Frechet Audio Distance from 1.40 to 1.22. Conditioning analyses further show that behavioral conditions improve target-anchor discriminability, while swapping behavior shifts generations toward the corresponding vocalization anchors. Demos are available on our project page.
Abstract (translated)
URL
https://arxiv.org/abs/2610.16382
PDF
-
WorldRoPE: Probabilistic World-Centric Rotary Embedding for Multi-View Attention
2026-10-12 07:56:46Mijeong Kim, Kyujin Lim, Bohyung HanAbstract
Rotary Positional Embedding (RoPE) is widely used to encode relative positions in Transformers. In multi-view settings, however, patch coordinates are defined independently in each image: the same 3D point can therefore have different coordinates across views, making them insufficient for establishing cross-view correspondences. We introduce WorldRoPE, a probabilistic world-centric RoPE for multi-view attention. At each layer, WorldRoPE back-projects each patch token into a set of volumetric Gaussian primitives in a shared world frame. These primitives represent multiple plausible 3D supports for a token, jointly capturing alternative depth hypotheses, depth uncertainty, and the finite spatial extent of its patch frustum. When a key token's support is projected into a query view, it induces a distribution over where that key may appear in the query image. WorldRoPE integrates the rotary operator over this distribution in closed form, combining rotations at the projected primitive centers according to their mixture weights with covariance-dependent attenuation. This makes attention less sensitive to uncertain key locations in the query view, without rasterizing or warping token features. The formulation reduces to deterministic RoPE when the support collapses to a point. Across layers, the predicted geometry guides cross-view attention, while the updated features refine the geometry used at the next layer. Experiments on novel view synthesis, depth estimation, and 3D geometry reconstruction demonstrate the effectiveness of WorldRoPE for multi-view Transformers.
Abstract (translated)
URL
https://arxiv.org/abs/2610.16337
PDF
