57 papers screened · since 2026-07-06 · updated July 13, 2026
The novel mechanism is a task-specific two-agent architecture utilizing confidence-calibrated answering and domain-specific numeric reasoning. Not a direct fit, but it's a higher-signal entry from the past week in Artificial Intelligence.
arXiv:2607.09623v1 ↗The novel mechanism is an incentive-compatible auction mechanism for dynamically allocating tasks to expert models and tools. Not a direct fit, but it's a higher-signal entry from the past week in Artificial Intelligence.
arXiv:2607.09600v1 ↗Recursive Evidence Replay addresses this by using model-internal relevance signals to construct a query-conditioned evidence pool. Not a direct fit, but it's a higher-signal entry from the past week in Artificial Intelligence.
arXiv:2607.02509v1 ↗The novel mechanism is the introduction of the vocabulary gap and the verifier gap, which are addressed through a unified framework of intelligence as cognitive discrepancy reduction. Not a direct fit, but it's a higher-signal entry from the past week in Machine Learning.
arXiv:2607.09560v1 ↗The novel mechanism is Latent Memory Palace (LMP), which formulates reasoning as variational inference with an autoregressive latent distribution. Not a direct fit, but it's a higher-signal entry from the past week in Machine Learning.
arXiv:2607.08724v1 ↗Evaluating vision-language models for safety-critical incidents in autonomous driving is challenging, AUTOPILOT-VQA addresses this gap with an incident-centric visual question answering benchmark, the dataset covers diverse safety-releva… Not a direct fit, but it's a higher-signal entry from the past week in Artificial Intelligence.
arXiv:2607.08745v1 ↗The Program-as-Weights (PAW) paradigm introduces a novel mechanism that compiles natural-language specifications into compact neural artifacts. Not a direct fit, but it's a higher-signal entry from the past week in Computation and Language (NLP).
arXiv:2607.02512v1 ↗The novel mechanism is a dual-channel debate framework that elicits both public and off-the-record responses. Not a direct fit, but it's a higher-signal entry from the past week in Computation and Language (NLP).
arXiv:2607.02507v1 ↗The core problem is enabling large models to reason through video generation, the novel mechanism is the Chain-of-Frame (CoF) reasoning technique, the headline empirical result is that Wan-CoF achieves considerable gains over the baselin… Not a direct fit, but it's a higher-signal entry from the past week in Artificial Intelligence.
arXiv:2607.08763v1 ↗The novel mechanism is Deep Gaussian Processes over DAGs, which places priors over functions to tackle these challenges. Not a direct fit, but it's a higher-signal entry from the past week in Statistical Machine Learning.
arXiv:2607.09645v1 ↗The Semantic Pareto-DQN proposes a multi-objective reinforcement learning framework that synthesizes transaction features into natural-language narratives. Not a direct fit, but it's a higher-signal entry from the past week in Machine Learning.
arXiv:2607.09641v1 ↗The novel LACUNA testbed introduces ground-truth parameter-level localization to evaluate unlearning methods. Not a direct fit, but it's a higher-signal entry from the past week in Computation and Language (NLP).
arXiv:2607.02513v1 ↗We built Notion's brief from a public sketch of what they do. Field Watch tunes to your actual roadmap and rivals — daily-fresh, delivered to your whole team.