The impressive performance of Large Language Models (LLMs) across various natural language processing tasks comes at the cost of vast computational resources and storage requirements. One-shot pruning techniques offer a way to alleviate these burdens by removing redundant weights without the need for retraining. Yet, the massive scale of LLMs often forces current pruning approaches to rely on heuristics instead of optimization-based techniques, potentially resulting in suboptimal compression. In this paper, we introduce ALPS, an optimization-based framework that tackles the pruning problem using the operator splitting technique and a preconditioned conjugate gradient-based post-processing step. Our approach incorporates novel techniques to accelerate and theoretically guarantee convergence while leveraging vectorization and GPU parallelism for efficiency. ALPS substantially outperforms state-of-the-art methods in terms of the pruning objective and perplexity reduction, particularly for highly sparse models. On the LLaMA3-8B model with 70\% sparsity, ALPS achieves a 29\% reduction in test perplexity on the WikiText dataset and a 8\% improvement in zero-shot benchmark performance compared to existing methods.

An elephant foot bone found by archaeologists digging in southern Spain may be evidence that a troop of war elephants stomped through ancient Europe. It would be the first concrete proof of the legendary Carthaginian General Hannibal's troop of battle elephants, according to academics. Drawings of Hannibal's war against the Romans had long suggested that the beasts were used in fighting, but no hard evidence backed up the theories. Now the creatures' skeletal remains appear to have been found in an Iron Age dig near Cordoba. Beyond ivory, the discovery of elephant remains in European archaeological contexts is exceptionally rare, says the team of scientists in a paper published in Journal of Archaeological Science: Reports.

One-shot pruning techniques offer a way to alleviate these burdens by removing redundant weights without the need for retraining. Y et, the massive scale of LLMs often forces current pruning approaches to rely on heuristics instead of optimization-based techniques, potentially resulting in suboptimal compression.

Recent workhasshown that neural networks can exhibit an inductive bias that is often introduced via designing specific structures[4].

One-shot pruning techniques offer a way to alleviate these burdens by removing redundant weights without the need for retraining. Y et, the massive scale of LLMs often forces current pruning approaches to rely on heuristics instead of optimization-based techniques, potentially resulting in suboptimal compression.

We consider the problem of estimating states ( e.g., position and velocity) and physical parameters ( e.g., friction, elasticity) from a sequence of observations

Existing work has shown that o1-level performance can be achieved with limited data distillation, but most existing methods focus on unidirectional supervised fine-tuning (SFT), overlooking the intricate interplay between diverse reasoning patterns. In this paper, we construct r1k, a high-quality reverse reasoning dataset derived by inverting 1,000 forward examples from s1k, and examine how SFT and Direct Preference Optimization (DPO) affect alignment under bidirectional reasoning objectives. SFT on r1k yields a 1.6%--6.8% accuracy improvement over s1k across evaluated benchmarks. However, naively mixing forward and reverse data during SFT weakens the directional distinction. Although DPO can partially recover this distinction, it also suppresses less preferred reasoning paths by shifting the probability mass toward irrelevant outputs. These findings suggest that mixed reasoning data introduce conflicting supervision signals, underscoring the need for robust and direction-aware alignment strategies.

In this paper, we are specifically interested in binary classification problems in the so-called streaming (or online) setting of active learning.

This paper documents GETALP's submission to the Third Run of the Automatic Minuting Shared Task at SIGDial 2025. We participated in Task B: question-answering based on meeting transcripts. Our method is based on a retrieval augmented generation (RAG) system and Abstract Meaning Representations (AMR). We propose three systems combining these two approaches. Our results show that incorporating AMR leads to high-quality responses for approximately 35% of the questions and provides notable improvements in answering questions that involve distinguishing between different participants (e.g., who questions).