中國 AI

該研究提出FSVL框架，結合輕量級ResNet50快速分類與YOLOv8n引導的慢速精細識別，並集成檢索增強的Qwen3-14B生成領域知識。在607張測試圖像上，快速分類器準確率達99.67%，級聯校正後達100%；知識生成任務中，科學名稱準確率從91.37%提升至100%， genus準確率從90.20%提升至98.82%，ROUGE-L從0.335提升至0.655。

一項針對Qwen3.5-27B工具使用任務的研究通過32個新場景發現，針對特定案例的殘差寫入干預能將熟悉的目標選擇正確率從約27%提升到約90%，但這些增益在同任務中重組熟悉事實和標識符時消失，且對候選資格或是否應選擇的獨立判斷改善甚微。

Yuval Kansal 等提出 Self-Direction 機制，讓智能體在任務失敗後從自身軌跡中提煉教訓，並即時蒸餾成僅約基礎模型參數量 0.02% 的低秩適配器，存入持久化經驗庫並按需路由。在 GAIA、Claw-Eval 和 Terminal-Bench 2.1 上，基於 Qwen 系列開放權重模型的智能體成功率提升 8.9 至 17.0 個百分點，重複錯誤顯著減少，少量適配器還能泛化到數十個新任務。

Hugging Face Bio 發佈 Carbon-X，一個基於 Qwen3-14B 的 14.8B 參數開源 DNA 語言模型，採用 Apache 2.0 許可。它可在對話中直接讀取原始 DNA 序列並用英文作答，在 5 項 DNA 基準上平均得分 66.7%，僅次於 GPT-6.1 Sol 的 67.5%。

斯坦福大學與加州大學伯克利分校的研究者發佈了 TokenGate，這是一個可插入現有大語言模型的學習式 Token 剪枝模塊。它通過一個小型 MLP 門控為輸入文檔中的每個 Token 打分並只保留得分最高的一部分，從而縮短長上下文預填充的首 Token 延遲（TTFT）1.4 至 11.3 倍，峯值顯存降低 1.5 至 3.4 倍，且不損失準確率。研究在 Qwen3-1.7B 與混合注意力的 Qwen3.5-2B 上驗證，代碼與檢查點以 Apache-2.0 許可開源。

Analytics Vidhya 對 Moonshot AI 的開放權重模型 Kimi K3.1 Pro 進行實測。該模型在多項基準上接近 Claude Opus 5.5，在金融文檔分析上領先，幻覺率較前代大幅下降，成本與生成速度優勢明顯；但在工具輔助 3D 任務與幻覺控制上仍落後於 Opus 5.5 和 GPT-6 Astra，完整權重將於 2026 年 11 月 4 日發佈，獨立驗證尚待進行。

Moonshot 推出 Kimi K3.1 Pro，這是一款2.8萬億總參數、1040億激活參數、支持百萬token上下文的多模態混合專家模型，在獨立基準測試中接近前沿水平。API 已在官網、應用和編程工具上線，官方承諾於2026年11月4日發佈開放權重。Artificial Analysis 將其排在智能指數第四位（55分），僅次於 Opus 5.5 的58分，開放與閉源之間的差距若權重如期放出將從12分縮小到3分。

Moonshot AI 推出 K3.1 系列首款模型 Kimi K3.1 Pro，採用 MoE 架構，總參數 2.8T、激活參數 104B，上下文窗口達 100 萬 token。它在 Terminal-Bench 4.0 上得分 65.7%，僅次於 Claude Opus 5.5；API 定價為每百萬 token 輸入 2 美元、輸出 10 美元，低於 K3。開放權重預計 2026 年 11 月釋出，目前尚不可自託管。

Moonshot 於 10 月 23 日將 Kimi K3.1 Pro 上線 Kimi API，目前尚未開放權重，但公司承諾在 2026 年 11 月 4 日前釋出完整模型權重。其定價為輸入每百萬 token 2 美元、輸出每百萬 token 10 美元，與 GPT-6.1 Sol 持平，約為 Claude Opus 5.5 的一半。Artificial Analysis 給出 55 分，僅次於 Opus 5.5 的 58 分，是該機構記錄中（即將）開放權重的模型首次如此接近前沿水平。

亞馬遜雲科技（AWS）宣佈 SageMaker AI 無服務器模型定製功能全面推出訓練會話（training sessions）。用户可自行編寫並協調後訓練循環，包括自定義損失函數、獎勵信號和多輪強化學習，而 GPU 分配、分佈式執行、故障恢復與檢查點存儲均由 SageMaker AI 託管。該功能支持 gpt-oss-20b、gemma-4-31B-it、Qwen3.6-27B 和 Qwen3.8-27B，按處理的 token 計費。

Liquid AI 推出專為實時請求路徑設計的推理引擎 RIPL，可在單張 B200 上以 50 毫秒以內完成端到端生成。官方基準顯示，其 p95 延遲比運行在 vLLM 上的 Qwen3.5 低 5.6 至 28.4 倍，Shopify 已將其用於生產搜索場景。不過 RIPL 僅面向企業客户、閉源，且所有數據均為自報。

該論文研究大型語言模型在處理法律敍事時，其內部激活中是否存在與CATO全部26個法律要素相對應的表徵。作者在Qwen3-1.7B、Qwen3-8B和Gemma-2 2B三個開放模型上進行線性可解碼性與稀疏自編碼器稀疏定位實驗，發現全部26個要素在三個模型中均可線性解碼，最佳層測試AUC為0.86至1.00，並報告了78個模型特定的層-特徵座標。

該論文提出ReAPO，一種從隨機最短路徑（Stochastic Shortest Paths）視角重新審視信用分配、由GRPO簡單修改而來的方法。它把額外生成預算分配給失敗軌跡，使成功的重試得到強化而不獎勵此前的失敗。在Qwen3與OLMo-2（0.6B至4B）以及八個數學與科學推理基準上，ReAPO的首輪嘗試表現優於GRPO和近期基線；在Qwen3-0.6B上，它以少39%和少31%的生成token分別達到GRPO雙倍組大小和雙倍訓練步數的峯值準確率。

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Train Qwen3 on AWS Trainium with torchtitan three ways on a single trn2.48xlarge: the Hugging Face Transformers v5 modeling backend, torchtitan's native Qwen3, and NKI mixture-of-experts kernels. The same torchrun launcher and config structure work unchanged across all three paths, including loading and saving published Hugging Face safetensors.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：An open source “bomb” drops from China

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：An open source “bomb” drops from China

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2610.27040v1 Announce Type: new Abstract: General-purpose multimodal models, including GPT-6 Astra and Claude Opus 5.5, demonstrate broad capabilities in embodied manipulation. Our analysis indicates that this generality is accompanied by strong spatial perception, as Astra can determine where to interact with task-relevant entities amid visual distractors. Because spatial perception connects high-level reasoning to the physical scene, we investigate scaling visual grounding to strengthen the perceptual foundation of general embodied operation. We develop a unified data pipeline that combines robot manipulation recordings with complementary visual data from diverse sources. After automatically annotating the robot recordings, the pipeline converts the col…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2610.26937v1 Announce Type: new Abstract: Large clinical language models often require multi-GPU infrastructure, limiting local deployment. We introduce CERTA, a decision-level statistical certificate that determines when a quantized model may substitute for its native BF16 reference and escalates decisions that cannot be certified to it. Across eleven instruction-tuned models, both general-purpose and domain-adapted, from three families (7-72B), we evaluated 2,500 clinical-verification decisions from de-identified patient contexts (2,484 patients), with patient-disjoint calibration and testing. In the higher-power operational analysis, round-to-nearest 8-bit quantization at BF16 storage served 99.8-100% of held-out decisions across all eleven models; for…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2610.26866v1 Announce Type: new Abstract: Mechanistic studies of large language models localize function in parameters, attention heads or neurons, but those parts do not specify how one inference unfolds. We treat the ordered layer-state trajectory as the empirical process and test whether layer order defines a measurable computation. Across representative Qwen, Llama and Gemma checkpoints, realized order produced paths 1.60--2.24 times as aligned with their endpoint chord as endpoint-preserving shuffles. The order effect transferred to a lexical task, but was stronger in three random initializations of the same Qwen architecture, identifying an architecture-compatible scaffold rather than a training-specific law. Direct hidden states provided stronger m…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2610.26887v1 Announce Type: new Abstract: Large language model (LLM) agents are evolving from conversational assistants into Cowork Agents capable of autonomous planning, tool use, and complex task execution. Financial analysis requires these agents to integrate cross-document evidence and perform multistep numerical reasoning, but expert task authoring is costly, while direct LLM synthesis struggles to produce challenging tasks with reliable answers. We introduce Navalia, a framework for Cowork Agents that combines verifiable task synthesis with long-horizon post-training. Using corporate financial reports as its data source, Navalia semantically selects supporting evidence from complete fact sets and constructs executable computation chains to generate…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Western open models strike back

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：DeepSeek has open-sourced the inference engine it uses to serve its models in production, a step it said last year The post DeepSeek open-sources the inference engine behind its API. And Huawei gets equal billing with Nvidia. appeared first on The New Stack.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：DeepSeek has open-sourced DeepInfer, the Apache 2.0 inference framework that serves DeepSeek-V4.1-Flash in production. It targets NVIDIA SM100 GPUs and Huawei Ascend 950 NPUs, and ships offline inference only. The post DeepSeek Open-Sources DeepInfer, the Inference Framework Behind DeepSeek-V4.1-Flash appeared first on MarkTechPost.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Use of tools ‘may result in the transfer of sensitive data to (China)’, defense department says The US government on Monday ordered its agencies to stop using artificial intelligence (AI) tools from Chinese startup DeepSeek, as the countries race for dominance in the fast-moving sector. A notice from the Department of Homeland Security (DHS) announced the ban on products developed by DeepSeek – and tools created using DeepSeek by third parties – saying they must be removed from IT devices. Continue reading...

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Congrats to DeepSeek V4.1 Flash 1017, MiniMax M3.1 Flash, and Mercury 3 (details below) but they’re not going to be in super wide use just yet.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Congrats to DeepSeek V4.1 Flash 1017, MiniMax M3.1 Flash, and Mercury 3 (details below) but they’re not going to be in super wide use just yet.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Multi-turn agentic workloads reprocess growing conversation context each turn, driving up latency. See how bi-directional KV cache transfer lets prefill and decode nodes reuse previously computed context across turns, cutting time to first token by up to 1.6x for Qwen3-32B and 3.3x for GLM-5-FP8 on Amazon EC2 P5en instances with EFA.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Not content with challenging the major frontier AI labs with lower-cost models, Chinese AI company DeepSeek is setting its sights The post DeepSeek made its name with cheap AI models. Now it’s undercutting search API rivals, too. appeared first on The New Stack.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2610.22454v1 Announce Type: new Abstract: Running prompt tokens through fewer routed experts during prefill is a simple way to make mixture-of-experts (MoE) serving cheaper, yet it is unclear what this damages. We locate the damage with token-, layer- and attention-path interventions on short-answer question answering in four MoE families. At half the experts the F1 loss of three families is mostly answer form: the model still finds the answer but stops answering briefly. In Qwen3-30B-A3B this loss is made at the two chat-template tokens before the answer, and in DeepSeek-V2-Lite at the first 8-32 prompt tokens, where its instruction sits; by answer likelihood these positions account for about half and a fifth of the damage. Reduction changes the first to…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2610.22381v1 Announce Type: new Abstract: When a fact becomes harder to elicit from a language model, has the model lost it, or has it merely become less accessible? Parameter-updating test-time continual learning agents are usually evaluated by current output accuracy; we ask whether that measurement captures how prior experience shapes later updates. We test whether learning a specific factual binding produces a later relearning advantage beyond familiarity with its entity, answer vocabulary, format, or query slot. In Qwen2.5-1.5B, we compare three matched, counterbalanced histories: LEARNED receives the correct binding, CONTROL receives matched exposure without it, and WRONG receives a competing relation-valid answer in the same slot. After an identica…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Post-train open LLMs like Qwen, GLM and Kimi into custom models you own, with SFT or RL. You write the training loop. Together runs the hard infrastructure.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：In the west the technology comes with at least some effort to provide security and a benign intent – not so from China There is much to fear from artificial intelligence developed in China and handed to businesses and governments in the west almost free of charge. Every month Chinese AI developers such as DeepSeek and Xiaomi dispatch the latest versions of their “open-weight” models that allow users to take the core software and use it without paying an expensive ongoing licence. Continue reading...

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯： Release: web-edit 0.1a2 web-edit (previously) is my tiny web app for browsing and editing files in a folder. I mainly built it for editing files on different machines via Tailscale. I added a "copy to clipboard" option for both text and image data. I needed this because I was running pelican prompts on a laptop from my phone (here's DeepSeek 4.1 Flash 1017) and I wanted an easy way to copy the resulting logs into a Gist. Tags: projects, pelican-riding-a-bicycle, deepseek

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：DeepSeek has just released DeepSeek-V4.1-Flash-1017, a new checkpoint of its V4.1-Flash model. It is live on the DeepSeek website, app and API. DeepSeek claims stronger scores, faster task completion and shorter outputs than the September V4.1-Flash release. The company calls it a Pareto improvement in performance and cost. Alongside it, DeepSeek opened its Search and […] The post DeepSeek Just Released V4.1-Flash-1017, Scores 77.0 on DeepSWE, Opens Search API appeared first on MarkTechPost.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2610.20627v1 Announce Type: new Abstract: Vision-language models are increasingly used for driving-scene understanding, yet their defensive reasoning under weakly structured Global South traffic remains unclear. We introduce BD-HazardVLM, a diagnostic benchmark of 500 Bangladesh road-scene images, comprising 375 from RSUD20K and 125 from TFP-BD, annotated for hazard presence, primary hazard category, risk timing, severity, defensive actions, visible evidence, and hallucination- and Western-road-bias traps. We evaluate Qwen2.5-VL-7B, InternVL3-8B, and MiniCPM-V-4.6 using a shared zero-shot task definition and structured output schema. Binary hazard-presence accuracy ranges from 88.0 percent to 89.2 percent, essentially matching the 89.0 percent majority-cl…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2610.20633v1 Announce Type: new Abstract: Long-horizon failures of language models are usually measured as end-task success or as a horizon length. We instead score, step by step, three components of a 400-step ledger task in which every step is easy in isolation and the model's own answers are re-injected into its context: recognising a record already seen 5 to 300 steps earlier, processing the current record locally, and writing a counter incremented at every step. On Qwen3-8B (non-thinking, 20 seeds, replicated on a second machine and backend), with every step scored component by component, two results hold. First, duplicate-recognition errors rise late in the episode -- from 0.0019 to 0.011 per step at risk across step 200 -- while the counter makes n…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2610.20681v1 Announce Type: new Abstract: Formal inequality proving requires mathematical reasoning and exact computation, often connected through intermediate lemmas and goal reformulations. We introduce IneqAgent, an agentic framework combining LLM reasoning with symbolic computation for inequality proving in Lean4. Lemma-level Monte Carlo Tree Search coordinates lemma construction, goal reformulation, and symbolic backend selection. Complementary sum-of-squares, Bernstein, and successive difference-substitution backends produce exact certificates for local Lean proof generation. The agent integrates these proofs with LLM-generated steps and uses Lean feedback to construct complete proofs checked against the original theorem statements. On four benchmar…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2610.19033v1 Announce Type: new Abstract: Low-rank adaptation (LoRA) applies every rank-one component of its update to every token, and mixture-of-experts variants replace this with a learned router that activates a fixed number of components. We argue that the number of components a token needs is a detection question with a classical answer. A token's responses to the rank-one components of a trained LoRA consist of a few large values on a Gaussian noise floor whose width matches that of an untrained adapter, and the floor carries cross-talk rather than adaptation. We introduce SQUELCH, a router-free mixture of rank-one experts that (i) soft-thresholds each token's responses at a multiple of its own robust noise level, so the active rank varies per toke…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2610.18883v1 Announce Type: new Abstract: Token-selection methods for on-policy distillation (OPD) prioritize student-visited positions using diverse signals, including uncertainty, teacher-student disagreement, and mismatch direction. We instead derive token importance from the distillation objective itself: how much can each position's loss decrease under an equal local KL budget on the student's predictive distribution? The answer is the utility $s_t=g_t^\top F_t^\dagger g_t$, where $g_t$ is the logit gradient and $F_t$ is the categorical Fisher matrix. For reverse-KL it equals the student-weighted variance of the student-teacher log-probability ratio, and for forward-KL it equals the Pearson $\chi^2$ divergence, so the appropriate utility follows from…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2610.18878v1 Announce Type: new Abstract: Post-training reasoning LLMs often requires making a tradeoff between costly supervision and compute inefficiency: supervised finetuning typically requires reasoning traces, which is data inefficient since in many applications one only has prompt-answer tuples and lacks the groundtruth chain-of-thought (CoT) sequence. On the other hand, reinforcement learning (with verifiable rewards, for example) can operate without CoT data but it is much more compute-inefficient since it requires multiple autoregressive completions per optimization step, none of which are guaranteed to give a non-zero policy gradient. In this work we explore a middleground, a self-distillation technique that uses answer-only data, treating CoT…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2610.14190v1 Announce Type: new Abstract: Explaining a vision--language model (VLM) prediction requires tracing training-example influence to the multimodal evidence that mediates it, yet existing attribution methods largely operate only at the sample level. Recent latent-mediated influence methods provide feature-level attribution in language models, but lack the multimodal and spatial structure needed to ground influence in VLMs. We introduce MOSAIC (Multimodal Orthogonalized Sparse Autoencoders for Influence Characterization). It learns a shared sparse image--text latent space whose concepts are 1) sparse within each modality, 2) supported on contiguous image regions, 3) aligned to the phrases that describe them, and 4) approximately decorrelated. It t…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2610.14118v1 Announce Type: new Abstract: The rapid advancement of Large Language Models (LLMs) has driven their deployment toward general-purpose harness environments, where LLM-based agents operate in sandbox workspaces and interact directly with file systems and shells. While numerous post-training pipelines exist for tool-use agents, their counterparts for harness-based agents remain scarce and are predominantly one-shot and teacher-dependent: a stronger external model synthesizes tasks and collects trajectories, the target agent is trained once, and evaluation results are consumed by human researchers rather than fed back into the loop. Such pipelines are inherently bounded by the capability of the external model and the scope of human prior, leaving…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：The social media giant tries to capitalize on, monetize its successful personal agent and maybe crack the business market. Chinese models go astray, like their U.S. counterparts.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：A new case of an out-of-control AI model shows that enterprises should treat models and agents as toddlers that need supervision and explicit instructions.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Training or finetuning a multi-billion parameter foundation model requires significant compute...

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯： Release: llm-gemini 0.35 I added support to llm-gemini for the new Gemini 4 Argon model, made available today via the Gemini API and for customers of their $100/month subscription plan. Here are the low, medium and high pelicans. As expected, they are all excellent: Gemini has a track record of producing great SVGs, and are the only lab where I know their staff have invested in SVG specifically (and who have boasted about their pelicans before.) Here's the "high" illustration, which cost me 17.166 cents: The pricing is interesting. At launch, Argon is $2/million input and $10/million output - the same price as GPT-6 Sol, half the price of Claude Opus 5.5 and 1/5th the price of Claude Fable 5.1 and GPT-6 Astra. But on December 1st Argon is set to double in pric…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Alibaba’s Qwen team has introduced QwenGyre, an elastic reinforcement learning RL framework for extreme-long-horizon agent training. QwenGyre is an online reinforcement learning system from Alibaba that trains coding agents through unmodified, black-box harnesses. It targets rollouts that run for hours and process close to 1M tokens. The QwenGyre elastic RL framework was used on Qwen’s […] The post Alibaba Qwen Introduces QwenGyre: Elastic Reinforcement Learning RL for XLong-Horizon Agents appeared first on MarkTechPost.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2610.12602v1 Announce Type: new Abstract: Continual learning (CL) seeks to adapt pretrained models while preserving prior capabilities. Yet the tasks supplied for preservation cover only part of what a model knows, while forgetting can extend beyond them. Recent work shows that language models can supplement this limited evidence by generating useful replay without access to their original training data. This raises a question about self-preservation: can a model help choose preservation targets that protect its own capabilities more broadly? To study this question, we organize task descriptions and their examples into Task Banks, whose entries can be selected and combined to guide replay generation by a frozen model. Continual instruction tuning with Qwe…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2610.12582v1 Announce Type: new Abstract: A latent recurrent language model applies a shared block of layers repeatedly, so one checkpoint can trade inference compute for accuracy by changing the loop count. At one loop, however, such models typically fall well short of a non-recurrent model of the same size, a one-loop tax that makes a separate non-recurrent model the better choice whenever latency matters. We present Elastic Latent Recurrence (ELR), a recipe for training looped language models without this tax. At one loop, an ELR model costs the same to serve as a non-recurrent model of the same size and, trained on the same data, reaches similar accuracy. Each additional loop then buys more accuracy for more inference compute. Three choices make this…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2610.12659v1 Announce Type: new Abstract: Data-centric scaling of RLVR has focused on collecting or generating verifiable training problems. However, corpus size alone does not determine whether the dataset provides a useful learning signal for a target policy. To guide the synthesis of effective RLVR data, we experimentally characterize dataset effectiveness as a function of its problems, the target policy, and the training rollout budget. Under this characterization, we show that a large open-source math corpus pooled from widely used RLVR datasets falls almost entirely into a zero-advantage trap: for a fixed target policy and rollout budget, roughly 88% of its problems are solved in every rollout or failed in every rollout. Filtering concentrates what…