发布日期：2026-04-21
收录条目：20
1. Silicon Valley has forgotten what normal people want
- 来源：The Verge AI
- 发布时间：2026-04-20 20:30 UTC
- 链接：https://www.theverge.com/tldr/915176/nft-metaverse-ai-weirdos
摘要：One of the most mortifying things about knowing a lot of techies is listening to them tell me excitedly about some very important discovery that they believe they have made. Recently, I ran into an acquaintance of mine,
2. Accelerate Generative AI Inference on Amazon SageMaker AI with G7e Instances
- 来源：AWS ML Blog
- 发布时间：2026-04-20 19:38 UTC
- 链接：https://aws.amazon.com/blogs/machine-learning/accelerate-generative-ai-inference-on-amazon-sagemaker-ai-with-g7e-instances/
摘要：Today, we are thrilled to announce the availability of G7e instances powered by NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs on Amazon SageMaker AI. You can provision nodes with 1, 2, 4, and 8 RTX PRO 6000 GPU insta
3. ToolSimulator: scalable tool testing for AI agents
- 来源：AWS ML Blog
- 发布时间：2026-04-20 17:06 UTC
- 链接：https://aws.amazon.com/blogs/machine-learning/toolsimulator-scalable-tool-testing-for-ai-agents/
摘要：You can use ToolSimulator, an LLM-powered tool simulation framework within Strands Evals, to thoroughly and safely test AI agents that rely on external tools, at scale. Instead of risking live API calls that expose perso
4. Fortnite developers can make AI characters now — just don’t try to date them
- 来源：The Verge AI
- 发布时间：2026-04-20 16:58 UTC
- 链接：https://www.theverge.com/games/914963/fortnite-ai-characters-developers-conversations
摘要：Following last year's AI-powered Darth Vader in Fortnite that swore in a re-creation of James Earl Jones' voice, Epic Games is now letting Fortnite creators experiment with a new "conversations" tool to create AI-powered
5. Omnichannel ordering with Amazon Bedrock AgentCore and Amazon Nova 2 Sonic
- 来源：AWS ML Blog
- 发布时间：2026-04-20 15:03 UTC
- 链接：https://aws.amazon.com/blogs/machine-learning/omnichannel-ordering-with-amazon-bedrock-agentcore-and-amazon-nova-2-sonic/
摘要：In this post, we'll show you how to build a complete omnichannel ordering system using Amazon Bedrock AgentCore, an agentic platform, to build, deploy, and operate highly effective AI agents securely at scale using any f
6. DeepER-Med: Advancing Deep Evidence-Based Research in Medicine Through Agentic AI
- 来源：arXiv cs.AI
- 发布时间：2026-04-20 04:00 UTC
- 链接：https://arxiv.org/abs/2604.15456
摘要：arXiv:2604.15456v1 Announce Type: new Abstract: Trustworthiness and transparency are essential for the clinical adoption of artificial intelligence (AI) in healthcare and biomedical research. Recent deep research systems
7. GIST: Multimodal Knowledge Extraction and Spatial Grounding via Intelligent Semantic Topology
- 来源：arXiv cs.AI
- 发布时间：2026-04-20 04:00 UTC
- 链接：https://arxiv.org/abs/2604.15495
摘要：arXiv:2604.15495v1 Announce Type: new Abstract: Navigating complex, densely packed environments like retail stores, warehouses, and hospitals poses a significant spatial grounding challenge for humans and embodied AI. In
8. Bureaucratic Silences: What the Canadian AI Register Reveals, Omits, and Obscures
- 来源：arXiv cs.AI
- 发布时间：2026-04-20 04:00 UTC
- 链接：https://arxiv.org/abs/2604.15514
摘要：arXiv:2604.15514v1 Announce Type: new Abstract: In November 2025, the Government of Canada operationalized its commitment to transparency by releasing its first Federal AI Register. In this paper, we argue that such regi
9. LACE: Lattice Attention for Cross-thread Exploration
- 来源：arXiv cs.AI
- 发布时间：2026-04-20 04:00 UTC
- 链接：https://arxiv.org/abs/2604.15529
摘要：arXiv:2604.15529v1 Announce Type: new Abstract: Current large language models reason in isolation. Although it is common to sample multiple reasoning paths in parallel, these trajectories do not interact, and often fail
10. Preregistered Belief Revision Contracts
- 来源：arXiv cs.AI
- 发布时间：2026-04-20 04:00 UTC
- 链接：https://arxiv.org/abs/2604.15558
摘要：arXiv:2604.15558v1 Announce Type: new Abstract: Deliberative multi-agent systems allow agents to exchange messages and revise beliefs over time. While this interaction is meant to improve performance, it can also create
11. Subliminal Transfer of Unsafe Behaviors in AI Agent Distillation
- 来源：arXiv cs.AI
- 发布时间：2026-04-20 04:00 UTC
- 链接：https://arxiv.org/abs/2604.15559
摘要：arXiv:2604.15559v1 Announce Type: new Abstract: Recent work on subliminal learning demonstrates that language models can transmit semantic traits through data that is semantically unrelated to those traits. However, it r
12. Bilevel Optimization of Agent Skills via Monte Carlo Tree Search
- 来源：arXiv cs.AI
- 发布时间：2026-04-20 04:00 UTC
- 链接：https://arxiv.org/abs/2604.15709
摘要：arXiv:2604.15709v1 Announce Type: new Abstract: Agent \texttt{skills} are structured collections of instructions, tools, and supporting resources that help large language model (LLM) agents perform particular classes of
13. The World Leaks the Future: Harness Evolution for Future Prediction Agents
- 来源：arXiv cs.AI
- 发布时间：2026-04-20 04:00 UTC
- 链接：https://arxiv.org/abs/2604.15719
摘要：arXiv:2604.15719v1 Announce Type: new Abstract: Many consequential decisions must be made before the relevant outcome is known. Such problems are commonly framed as \emph{future prediction}, where an LLM agent must form
14. LLM Reasoning Is Latent, Not the Chain of Thought
- 来源：arXiv cs.AI
- 发布时间：2026-04-20 04:00 UTC
- 链接：https://arxiv.org/abs/2604.15726
摘要：arXiv:2604.15726v1 Announce Type: new Abstract: This position paper argues that large language model (LLM) reasoning should be studied as latent-state trajectory formation rather than as faithful surface chain-of-thought
15. Structured Abductive-Deductive-Inductive Reasoning for LLMs via Algebraic Invariants
- 来源：arXiv cs.AI
- 发布时间：2026-04-20 04:00 UTC
- 链接：https://arxiv.org/abs/2604.15727
摘要：arXiv:2604.15727v1 Announce Type: new Abstract: Large language models exhibit systematic limitations in structured logical reasoning: they conflate hypothesis generation with verification, cannot distinguish conjecture f
16. KWBench: Measuring Unprompted Problem Recognition in Knowledge Work
- 来源：arXiv cs.AI
- 发布时间：2026-04-20 04:00 UTC
- 链接：https://arxiv.org/abs/2604.15760
摘要：arXiv:2604.15760v1 Announce Type: new Abstract: We introduce the first version of KWBench (Knowledge Work Bench), a benchmark for unprompted problem recognition in large language models: can an LLM identify a professiona
17. Stein Variational Black-Box Combinatorial Optimization
- 来源：arXiv cs.AI
- 发布时间：2026-04-20 04:00 UTC
- 链接：https://arxiv.org/abs/2604.15837
摘要：arXiv:2604.15837v1 Announce Type: new Abstract: Combinatorial black-box optimization in high-dimensional settings demands a careful trade-off between exploiting promising regions of the search space and preserving suffic
18. Discover and Prove: An Open-source Agentic Framework for Hard Mode Automated Theorem Proving in Lean 4
- 来源：arXiv cs.AI
- 发布时间：2026-04-20 04:00 UTC
- 链接：https://arxiv.org/abs/2604.15839
摘要：arXiv:2604.15839v1 Announce Type: new Abstract: Most ATP benchmarks embed the final answer within the formal statement -- a convention we call "Easy Mode" -- a design that simplifies the task relative to what human compe
19. Experience Compression Spectrum: Unifying Memory, Skills, and Rules in LLM Agents
- 来源：arXiv cs.AI
- 发布时间：2026-04-20 04:00 UTC
- 链接：https://arxiv.org/abs/2604.15877
摘要：arXiv:2604.15877v1 Announce Type: new Abstract: As LLM agents scale to long-horizon, multi-session deployments, efficiently managing accumulated experience becomes a critical bottleneck. Agent memory systems and agent sk
20. Towards Rigorous Explainability by Feature Attribution
- 来源：arXiv cs.AI
- 发布时间：2026-04-20 04:00 UTC
- 链接：https://arxiv.org/abs/2604.15898
摘要：arXiv:2604.15898v1 Announce Type: new Abstract: For around a decade, non-symbolic methods have been the option of choice when explaining complex machine learning (ML) models. Unfortunately, such methods lack rigor and ca