Quantization, distillation, pruning and serving work aimed at the same accuracy for less memory, latency and money.
arXiv:2608. 08638v1 Announce Type: cross Abstract: Zero-shot text-to-speech (TTS) now supports interactive assistants, personalized media, and accessibility tools.
arXiv:2608. 09225v1 Announce Type: cross Abstract: The key-value (KV) cache is the primary throughput optimization in modern large language model (LLM) inference, enabling prefix reuse across requests.
arXiv:2608. 09109v1 Announce Type: new Abstract: User feedback offers natural supervision for persistent LLM improvement, but a single message may support multiple behavioral changes with different scopes of generalization.
arXiv:2608. 09637v1 Announce Type: cross Abstract: Diffusion models have enabled high-quality video generation in recent years, but the high cost of iterative sampling hinders their practical deployment.
arXiv:2608. 09826v1 Announce Type: cross Abstract: Reinforcement learning with verifiable rewards yields no group-relative signal when rollout groups are uniformly correct or uniformly wrong, which account for 63.
arXiv:2608. 08569v1 Announce Type: new Abstract: Recent advancements in Speech Large Language Models have demonstrated remarkable capabilities in understanding complex audio tasks.
arXiv:2608. 09444v1 Announce Type: new Abstract: A main promise of looped language models (LMs) is depth-adaptive inference.
arXiv:2608. 08389v1 Announce Type: new Abstract: Long-horizon research agents solve open-ended tasks through iterative retrieval, aggregation, and synthesis, but context grows rapidly while the marginal value of additional evidence often declines.
Free-viewpoint 3D scene media is increasingly important for immersive applications, yet practical capture often suffers from severe view sparsity and motion blur. Although neural rendering has advanced sparse-view synthesis, existing blur-aware methods typically require substantial multi-view redundancy, accurate camera poses, or costly per-scene optimization.
We present TRACE-GS, an on-policy trajectory distillation framework that leverages privileged geometric conditioning at training time, thereby adapting a diffusion prior to sparse-view 3D Gaussian Splatting (3DGS) restoration. Rather than pursuing increasingly sophisticated restoration architectures, we identify a more fundamental limitation shared by existing diffusion-based approaches: supervision at independently noised states does not cover those reached during inference.
Self-improvement for multimodal large language models (MLLMs) is typically driven by reward-based methods that provide only coarse scalar feedback. Distillation offers a richer alternative through dense token-level supervision, but in the visual domain it usually depends on privileged context constructed using external annotations and tools, or stronger models.
Reinforcement learning with verifiable rewards yields no group-relative signal when rollout groups are uniformly correct or uniformly wrong, which account for 63. 0-68.
Industrial anomaly detection (IAD) requires identifying fine-grained deviations from normal visual patterns. Multimodal large language models (MLLMs) can improve recognition accuracy by comparing query images with references at inference time, but these benefits rely on additional retrieval and processing.
Diffusion models have enabled high-quality video generation in recent years, but the high cost of iterative sampling hinders their practical deployment. Few-step distillation alleviates this cost, yet exposes a quality--diversity trade-off between its two dominant paradigms: trajectory-level distillation (e.
Omnimodal language models (OLMs) enable unified audio-visual understanding, but processing long joint token sequences makes inference computationally prohibitive. While recent token compression methods attempt to alleviate this burden, compressing modalities in isolation often destroys the temporal cross-modal anchors necessary for coherent reasoning.
Efficient text-to-image generation requires both reinforcement-learning (RL)-based reward alignment and few-step distillation, yet these procedures are typically performed sequentially, increasing training cost and risking the loss of reward gains during compression. We instead take an RL-native perspective: diffusion RL already generates reward-scored finite-step trajectories, whose intermediate states provide a natural source of distillation supervision rather than a disposable byproduct of sampling.
We introduce Motif 3, a decoder-only Mixture-of-Experts language model with 314 billion total parameters and 13. 2 billion activated per token.
arXiv:2608. 07193v1 Announce Type: new Abstract: Visual-token pruning can substantially reduce the inference cost of multimodal large language models (MLLMs), yet existing methods largely rely on fixed, handcrafted heuristics and costly expert trial and error.
arXiv:2608. 07088v1 Announce Type: cross Abstract: Multimodal large language models (MLLMs) encode images as long visual token sequences, making prefilling and KV-cache storage expensive.