Quantization, distillation, pruning and serving work aimed at the same accuracy for less memory, latency and money.
arXiv:2606. 09012v1 Announce Type: cross Abstract: Post-training quantization (PTQ) converts a trained full-precision model into low-bit weights without task-level retraining, while quantization-aware training (QAT) incorporates quantization into the training loop.
arXiv:2605. 15491v2 Announce Type: replace-cross Abstract: Layer pruning removes entire Transformer decoder blocks from large language models, but introduces a mismatch between the hidden state received by the next surviving layer and the distribution it was trained to process, leading to significant performance degradation.
arXiv:2606. 09091v1 Announce Type: new Abstract: On-policy distillation (OPD) has recently emerged as an important post-training paradigm.
arXiv:2606. 07878v1 Announce Type: new Abstract: The KV cache is the memory bottleneck of long-horizon language model deployment.
arXiv:2603. 05500v2 Announce Type: replace-cross Abstract: Efficient and stable training of large language models (LLMs) remains a core challenge in modern machine learning systems.
arXiv:2606. 08696v1 Announce Type: cross Abstract: Counterfactual recourse aims to provide actionable feature changes that would alter an unfavorable decision made by a predictive model.
arXiv:2606. 08850v1 Announce Type: cross Abstract: Inference-Time Scaling (ITS) has largely succeeded in verifiable domains like math and coding, where cheap verification enables scalable output selection.
arXiv:2606. 07836v1 Announce Type: cross Abstract: Many-body GW-Bethe-Salpeter equation calculations are essential for accurate simulations of electronic structure and optical properties in modern low-dimensional nanomaterials.
Video world models that maintain 3D spatial consistency across generated frames typically rely on explicit point cloud memory constructed in RGB space. This design is both computationally expensive, requiring repeated rendering and VAE encoding, and inherently lossy, as the round trip through pixel space discards rich features of the learned latent representation.
Long-context language model inference is bottlenecked by memory, as the KV cache grows with context length. Recent techniques to compress the KV cache fall short: they either degrade model quality substantially or require considerable time and compute to compress a single long prompt.
3D object detection is the backbone of perception for automated vehicles (AV) and broader intelligent transportation systems applications. Long-range detection is challenging because sensing evidence is sparse; yet this ``long-range'' scenario is routine in traffic.
3D semantic scene generation is crucial for autonomous driving applications, yet most methods rely on complex 3D-specific architectures such as triplane encoders and adapted diffusion networks, limiting both their simplicity and their editing capabilities. We propose EditSSC, an editing-ready method for 3D semantic scene generation using 2D Bird's Eye View (BEV) representations and off-the-shelf latent diffusion network.
arXiv:2605. 25054v2 Announce Type: replace-cross Abstract: Deploying deep neural networks on resource-constrained 6G edge devices demands aggressive compression with minimal accuracy loss.
arXiv:2606. 03002v2 Announce Type: replace-cross Abstract: Quantization is a standard path to deploying large language models, and quantized models are typically judged acceptable when perplexity or downstream accuracy remains close to the full-precision original.
arXiv:2606. 07491v1 Announce Type: cross Abstract: High-performance computing (HPC) clusters remain the backbone of large-scale scientific computation, traditionally executing deterministic, linear pipelines optimised for predictable performance.
arXiv:2606. 07082v1 Announce Type: cross Abstract: On-policy distillation (OPD) is increasingly used to improve large language model reasoning, but its training dynamics remain poorly understood.
arXiv:2606. 06566v1 Announce Type: cross Abstract: Agentic tool-calling language models depend on large registries of callable APIs, functions, and local actions.
arXiv:2606. 06943v1 Announce Type: cross Abstract: Vision-language models (VLMs) such as CLIP achieve strong zero-shot recognition but remain highly fragile under adversarial perturbations.
arXiv:2603. 20967v2 Announce Type: replace-cross Abstract: One of the most common machine learning setups is logistic regression.
arXiv:2605. 08692v2 Announce Type: replace Abstract: Post-training weight-only quantization to 4 bits is widely used to reduce the memory and compute costs of large language model inference.