{"version":"https://jsonfeed.org/version/1.1","title":"Apple — Lab Ledger Desk","home_page_url":"https://labledgerdesk.pages.dev/lab/apple/","feed_url":"https://labledgerdesk.pages.dev/lab/apple/feed.json","description":"Official Apple announcements filed by Lab Ledger Desk.","items":[{"id":"https://labledgerdesk.pages.dev/b/2026/11/12/theoretical-limits-of-language-model-alignment/","url":"https://labledgerdesk.pages.dev/b/2026/11/12/theoretical-limits-of-language-model-alignment/","title":"Theoretical Limits of Language Model Alignment","content_text":"Language model (LM) alignment improves model outputs to reflect human preferences while preserving the capabilities of the base model. …","date_published":"2026-11-12T00:00:00.000Z","authors":[{"name":"Apple"}],"tags":["note","llm","safety"],"external_url":"https://machinelearning.apple.com/research/theoretical-limits-alignment"},{"id":"https://labledgerdesk.pages.dev/b/2026/11/10/evaluating-the-evaluators-investigating-llm-judges-for-personalized/","url":"https://labledgerdesk.pages.dev/b/2026/11/10/evaluating-the-evaluators-investigating-llm-judges-for-personalized/","title":"Evaluating the Evaluators: Investigating LLM Judges for Personalized Writing Style Assessment","content_text":"This paper was accepted at the Can We Trust AI Evaluation? Workshop at NeurIPS 2026. We evaluate how good large language model (LLM) judges are for personalized writing tasks. …","date_published":"2026-11-10T00:00:00.000Z","authors":[{"name":"Apple"}],"tags":["research","llm"],"external_url":"https://machinelearning.apple.com/research/evaluating-evaluators-llm-judges"},{"id":"https://labledgerdesk.pages.dev/b/2026/11/10/on-the-importance-of-gating-memorization-vs-in-context-learning-in/","url":"https://labledgerdesk.pages.dev/b/2026/11/10/on-the-importance-of-gating-memorization-vs-in-context-learning-in/","title":"On the Importance of Gating: Memorization vs. In-Context Learning in State Space Models","content_text":"State Space Models (SSMs) have emerged as a compelling alternative to Transformers, enabling sequence modeling with constant memory and linear compute. …","date_published":"2026-11-10T00:00:00.000Z","authors":[{"name":"Apple"}],"tags":["note"],"external_url":"https://machinelearning.apple.com/research/gating-memorization-in-context"},{"id":"https://labledgerdesk.pages.dev/b/2026/11/10/weighted-state-souping-for-efficient-long-context-retrieval/","url":"https://labledgerdesk.pages.dev/b/2026/11/10/weighted-state-souping-for-efficient-long-context-retrieval/","title":"Weighted State Souping for Efficient Long-Context Retrieval","content_text":"This paper was accepted at the Workshop on Personalized, Aligned, Long-term Memory for AI Systems (PALM) at NeurIPS 2026. State space models (SSMs) combine fast inference and fixed state size with competitive language-modeling performance. …","date_published":"2026-11-10T00:00:00.000Z","authors":[{"name":"Apple"}],"tags":["research","llm"],"external_url":"https://machinelearning.apple.com/research/weighted-state-souping"},{"id":"https://labledgerdesk.pages.dev/b/2026/11/05/faster-retrieval-for-better-more-efficient-rag/","url":"https://labledgerdesk.pages.dev/b/2026/11/05/faster-retrieval-for-better-more-efficient-rag/","title":"Faster Retrieval for Better, More Efficient RAG","content_text":"Maximum inner product search (MIPS) sits at the core of recommendation systems, search, and retrieval-augmented generation (RAG). Given a query vector, MIPS systems search within a database of millions of candidate key vectors for the one that best matches the query. …","date_published":"2026-11-05T00:00:00.000Z","authors":[{"name":"Apple"}],"tags":["note"],"external_url":"https://machinelearning.apple.com/research/faster-retrieval"},{"id":"https://labledgerdesk.pages.dev/b/2026/11/03/attention-to-mamba-a-recipe-for-cross-architecture-distillation/","url":"https://labledgerdesk.pages.dev/b/2026/11/03/attention-to-mamba-a-recipe-for-cross-architecture-distillation/","title":"Attention to Mamba: A Recipe for Cross-Architecture Distillation","content_text":"State Space Models (SSMs) such as Mamba have become a popular alternative to Transformer models, due to their reduced memory consumption and higher throughput at generation compared to their Attention-based counterparts. …","date_published":"2026-11-03T00:00:00.000Z","authors":[{"name":"Apple"}],"tags":["note"],"external_url":"https://machinelearning.apple.com/research/attention-mamba-distillation"},{"id":"https://labledgerdesk.pages.dev/b/2026/11/02/online-dense-reinforcement-learning-with-planning-grounding-agent-for/","url":"https://labledgerdesk.pages.dev/b/2026/11/02/online-dense-reinforcement-learning-with-planning-grounding-agent-for/","title":"Online Dense Reinforcement Learning with Planning-Grounding Agent for UI Navigation on Live iOS Apps","content_text":"While frontier vision-language models demonstrate strong capabilities in user interface (UI) navigation, their high inference costs prohibit everyday use. …","date_published":"2026-11-02T00:00:00.000Z","authors":[{"name":"Apple"}],"tags":["note","llm","agents"],"external_url":"https://machinelearning.apple.com/research/planning-grounding-ui-navigation"}]}