{
    "ok": true,
    "version": "3.0.0-mission-control",
    "checkedAt": "2026-09-12T20:32:46+00:00",
    "elapsedMs": 125,
    "count": 20,
    "categories": {
        "ai": 20
    },
    "items": [
        {
            "id": "02891c1423845b6d4d46022626679da972f4f233",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "Probabilistic Focal Search: Accelerating Bounded-Suboptimal Search via Lower-Bound Advancement",
            "summary": "arXiv:2609.10584v1 Announce Type: new Abstract: Bounded-suboptimal search seeks a solution within a factor $w$ of optimal while reducing search effort. Focal Search (FS) uses heuristic guidance within FOCAL, the frontier nodes eligible under the threshold $w f_{\\min}$, but its deterministic policy may leave $f_{\\min}$ unchanged for many expansions. We introduce Probabilistic Focal Search (PFS), which follows the FS guided choice with probability $p$ and expands a minimum-$f$ OPEN node with probability $1-p$. The latter branch encourages the lower bound to advance, enlarging FOCAL and admitting nodes that may lead to feasible solutions. By balancing guidance and lower-bound advancement, this mechanism can reduce time to a bounded solution when progress is limited by delayed FOCAL admission. As a secondary transfer experiment, we apply the same scheduler to Dynamic Potential Search, yield…",
            "url": "https://arxiv.org/abs/2609.10584",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "Probabilistic Focal Search: Accelerating Bounded-Suboptimal Search via Lower-Bound Advancement",
                "summary": "arXiv:2609.10584v1 Announce Type: new \nAbstract: Bounded-suboptimal search seeks a solution within a factor $w$ of optimal while reducing search effort. Focal Search (FS) uses heuristic guidance within FOCAL, the frontier nodes eligible under the threshold $w f_{\\min}$, but its deterministic policy may leave $f_{\\min}$ unchanged for many expansions. We introduce Probabilistic Focal Search (PFS), which follows the FS guided choice with probability $p$ and expands a minimum-$f$ OPEN node with probability $1-p$. The latter branch encourages the lower bound to advance, enlarging FOCAL and admitting nodes that may lead to feasible solutions. By balancing guidance and lower-bound advancement, this mechanism can reduce time to a bounded solution when progress is limited by delayed FOCAL admission. As a secondary transfer experiment, we apply the same scheduler to Dynamic Potential Search, yielding Probabilistic Dynamic Potential Search (PDPS). We benchmark PFS against FS on N-Puzzle, Pancake Sorting, and the Traveling Salesperson Problem (TSP), and evaluate its anytime extension on the Generalized Covering TSP (GCTSP), using multiple $w$ and $p$ values. Across these benchmarks, the largest gains occur when long $f_{\\min}$ plateaus delay useful FOCAL admissions; in such settings, the probabilistic factor may reduce node expansions by about 90\\% or more (e.g., on N-Puzzle and TSP). For the anytime algorithm family, Anytime Probabilistic Focal Search (APFS) outperforms all tested algorithms in evaluating anytime methods on GCTSP. We also observe that the benefit is smaller when the deterministic search already advances efficiently (e.g., Pancake Sorting), indicating that the probabilistic factor is most useful when FOCAL admission is a search bottleneck. The PDPS transfer shows that the mechanism also transfers to potential guidance, although its common-success effects remain domain- and bound-dependent.",
                "url": "https://arxiv.org/abs/2609.10584",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        },
        {
            "id": "fe113aa3b58b1cfa7ae0a06882dc645b221b80ce",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "Automating Quadratic Unconstrained Binary Optimization (QUBO) Formulation Generation from Natural Language",
            "summary": "arXiv:2609.10629v1 Announce Type: new Abstract: Quadratic Unconstrained Binary Optimization (QUBO) is a central formulation for combinatorial optimization and has gained increasing attention due to its compatibility with quantum, hybrid quantum-classical, and quantum-inspired solvers. However, translating natural-language problem descriptions into correct QUBO formulations remains difficult, requiring the identification of binary variables, constraints, objective functions, penalty terms, and suitable penalty weights. This process is time-consuming and often demands substantial domain expertise. To address this challenge, we propose an end-to-end multi-agent framework that automatically generates QUBO formulations from natural-language problem descriptions, supported by structured or unstructured test cases. To evaluate its performance, We also introduce QUBOBench, a benchmark containin…",
            "url": "https://arxiv.org/abs/2609.10629",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "Automating Quadratic Unconstrained Binary Optimization (QUBO) Formulation Generation from Natural Language",
                "summary": "arXiv:2609.10629v1 Announce Type: new \nAbstract: Quadratic Unconstrained Binary Optimization (QUBO) is a central formulation for combinatorial optimization and has gained increasing attention due to its compatibility with quantum, hybrid quantum-classical, and quantum-inspired solvers. However, translating natural-language problem descriptions into correct QUBO formulations remains difficult, requiring the identification of binary variables, constraints, objective functions, penalty terms, and suitable penalty weights. This process is time-consuming and often demands substantial domain expertise. To address this challenge, we propose an end-to-end multi-agent framework that automatically generates QUBO formulations from natural-language problem descriptions, supported by structured or unstructured test cases. To evaluate its performance, We also introduce QUBOBench, a benchmark containing 100 combinatorial optimization problems across 12 application domains, curated from peer-reviewed literature, competitions, and canonical NP-hard problems. Experimental results show that our framework achieves 68% accuracy on QUBOBench, outperforming a direct single-call baseline by 22%. Further analysis identifies iterative self-repair as the most important component contributing to improved performance. The data and code are open-sourced at https://quitttcat.github.io/QuantumQUBOAgent.",
                "url": "https://arxiv.org/abs/2609.10629",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        },
        {
            "id": "2d9811f68a9e10222b223f905e23f34873998a8e",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "A Multi-Stage Rule-Chaining Framework for Compositional and Interpretable Cognitive Reasoning",
            "summary": "arXiv:2609.10654v1 Announce Type: new Abstract: The Abstraction and Reasoning Corpus (ARC) benchmarks cognitive generalization, the ability to infer and apply abstract rules from limited examples. This paper presents a multi-stage rule-chaining framework that performs compositional reasoning across symbolic, structural, and conceptual levels. The framework integrates three complementary solvers: (1) a deterministic rule discovery module that induces atomic transformations through geometric, color, and object-based analysis; (2) a pattern-composition engine that reconstructs outputs via block merging, repetition, and spatial heuristics; and (3) a structural abstraction layer that infers hierarchical and nested relationships across grids. These solvers operate sequentially within a progressive fallback hierarchy, where each stage reuses prior reasoning traces to enhance interpretability a…",
            "url": "https://arxiv.org/abs/2609.10654",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "A Multi-Stage Rule-Chaining Framework for Compositional and Interpretable Cognitive Reasoning",
                "summary": "arXiv:2609.10654v1 Announce Type: new \nAbstract: The Abstraction and Reasoning Corpus (ARC) benchmarks cognitive generalization, the ability to infer and apply abstract rules from limited examples. This paper presents a multi-stage rule-chaining framework that performs compositional reasoning across symbolic, structural, and conceptual levels. The framework integrates three complementary solvers:\n  (1) a deterministic rule discovery module that induces atomic transformations through geometric, color, and object-based analysis;\n  (2) a pattern-composition engine that reconstructs outputs via block merging, repetition, and spatial heuristics; and\n  (3) a structural abstraction layer that infers hierarchical and nested relationships across grids.\n  These solvers operate sequentially within a progressive fallback hierarchy, where each stage reuses prior reasoning traces to enhance interpretability and generalization. Training passed for 995 tasks out of 1000, further evaluated on 105 tasks out of 120 and solved 230 test tasks out of 240 ARC-AGI-2 tasks. The system achieved strong coverage across deterministic, compositional, and abstract categories, demonstrating an overall accuracy exceeding 95 percent. The proposed architecture bridges symbolic reasoning and pattern synthesis, providing interpretable insight into cognitive generalization. The results suggest that rule chaining and hierarchical composition can advance machine reasoning toward transparent, human-aligned abstraction without relying on task-specific tuning.",
                "url": "https://arxiv.org/abs/2609.10654",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        },
        {
            "id": "6d0ad7fc3fcc462c8bbb70a17cb3af0c1168e70d",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "Understanding LoRA Rank Trade-offs in Diffusion Model Fine-Tuning",
            "summary": "arXiv:2609.10656v1 Announce Type: new Abstract: Selecting LoRA rank for diffusion fine-tuning requires balancing quality and compute cost. We present a controlled study on CIFAR-10 using a DDPM U-Net with ranks {2,4,8,16,32}, fixed optimization settings, and a reproducible local-folder pytorch-fid protocol. We report FID, trainable parameters, runtime, and GPU memory, then validate trends with extended-budget DDPM runs (20 epochs; ranks 4/8/16) and a Tiny DiT backbone (10 epochs; ranks 4/8/16). Results show moderate ranks are most efficient: rank 4 achieves the best DDPM FID (124.1380), rank 8 is close (124.2136), and higher ranks provide limited gains despite larger adaptation cost. These findings support small-to-moderate ranks as practical defaults under fixed training budgets.",
            "url": "https://arxiv.org/abs/2609.10656",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "Understanding LoRA Rank Trade-offs in Diffusion Model Fine-Tuning",
                "summary": "arXiv:2609.10656v1 Announce Type: new \nAbstract: Selecting LoRA rank for diffusion fine-tuning requires balancing quality and compute cost. We present a controlled study on CIFAR-10 using a DDPM U-Net with ranks {2,4,8,16,32}, fixed optimization settings, and a reproducible local-folder pytorch-fid protocol. We report FID, trainable parameters, runtime, and GPU memory, then validate trends with extended-budget DDPM runs (20 epochs; ranks 4/8/16) and a Tiny DiT backbone (10 epochs; ranks 4/8/16). Results show moderate ranks are most efficient: rank 4 achieves the best DDPM FID (124.1380), rank 8 is close (124.2136), and higher ranks provide limited gains despite larger adaptation cost. These findings support small-to-moderate ranks as practical defaults under fixed training budgets.",
                "url": "https://arxiv.org/abs/2609.10656",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        },
        {
            "id": "56b5190a87aa9cc5b5a355e64032fc43de7d3eff",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "Quantifying the Memorization-to-Generalization Transition: Scaling Laws and Phase Structure in Grokking",
            "summary": "arXiv:2609.10657v1 Announce Type: new Abstract: Neural networks trained past memorization frequently undergo a delayed transition to generalization, a phenomenon known as grokking. Despite theoretical progress on \\emph{why} this transition occurs, the quantitative structure of \\emph{when} it occurs in hyperparameter space remains uncharacterized. We map the memorization-to-generalization boundary across 384 configurations of two-hidden-layer MLPs on modular arithmetic, fitting a power-law scaling relation for generalization onset time: $T_{\\mathrm{grok}} \\propto H^{-0.27}\\, D^{-2.04}\\, \\eta^{-0.50}\\, \\lambda^{-0.64}$ ($R^2 = 0.732$; $0.821$ with interactions). The exponent hierarchy reveals that data complexity ($D^{-2.04}$) is the dominant driver of regime transition, not model capacity ($H^{-0.27}$): doubling data accelerates generalization by ${\\sim}4\\times$, while doubling width yie…",
            "url": "https://arxiv.org/abs/2609.10657",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "Quantifying the Memorization-to-Generalization Transition: Scaling Laws and Phase Structure in Grokking",
                "summary": "arXiv:2609.10657v1 Announce Type: new \nAbstract: Neural networks trained past memorization frequently undergo a delayed transition to generalization, a phenomenon known as grokking. Despite theoretical progress on \\emph{why} this transition occurs, the quantitative structure of \\emph{when} it occurs in hyperparameter space remains uncharacterized. We map the memorization-to-generalization boundary across 384 configurations of two-hidden-layer MLPs on modular arithmetic, fitting a power-law scaling relation for generalization onset time: $T_{\\mathrm{grok}} \\propto H^{-0.27}\\, D^{-2.04}\\, \\eta^{-0.50}\\, \\lambda^{-0.64}$ ($R^2 = 0.732$; $0.821$ with interactions). The exponent hierarchy reveals that data complexity ($D^{-2.04}$) is the dominant driver of regime transition, not model capacity ($H^{-0.27}$): doubling data accelerates generalization by ${\\sim}4\\times$, while doubling width yields only ${\\sim}1.2\\times$. A sharp phase boundary at weight decay $\\lambda \\gtrsim 1.0$ separates grokking from non-grokking configurations, and weight norm trajectories show monotonic compression during the transition, consistent with implicit regularization selecting low-complexity solutions. These results provide a quantitative foundation for predicting and controlling regime transitions in overparameterized networks.",
                "url": "https://arxiv.org/abs/2609.10657",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        },
        {
            "id": "67ac1c768235849522525ffe6c194eaf9032c63b",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "An Open Recipe for IMO Gold: Training Nemotron for Olympiad Mathematics",
            "summary": "arXiv:2609.10712v1 Announce Type: new Abstract: We study how model post-training and test-time inference design affect natural-language proof generation for hard olympiad mathematics. Starting from Nemotron 3 Ultra, we train two specialist checkpoints using supervised fine-tuning and reinforcement learning, and evaluate checkpoint choice, verification, and refinement. Based on these findings, we present an open-model test-time-compute pipeline. The system operates entirely in natural language, with no formal prover, external tools, or internet access. Three Nemotron 3 Ultra checkpoints - the general-availability model and two post-trained specialists - power an iterative search that generates, verifies, and refines candidate proofs; a separate high-compute stage then selects each final submission. The system scored 30 out of 42 points at IMO 2026, reaching the gold-medal threshold. We r…",
            "url": "https://arxiv.org/abs/2609.10712",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "An Open Recipe for IMO Gold: Training Nemotron for Olympiad Mathematics",
                "summary": "arXiv:2609.10712v1 Announce Type: new \nAbstract: We study how model post-training and test-time inference design affect natural-language proof generation for hard olympiad mathematics. Starting from Nemotron 3 Ultra, we train two specialist checkpoints using supervised fine-tuning and reinforcement learning, and evaluate checkpoint choice, verification, and refinement. Based on these findings, we present an open-model test-time-compute pipeline. The system operates entirely in natural language, with no formal prover, external tools, or internet access. Three Nemotron 3 Ultra checkpoints - the general-availability model and two post-trained specialists - power an iterative search that generates, verifies, and refines candidate proofs; a separate high-compute stage then selects each final submission. The system scored 30 out of 42 points at IMO 2026, reaching the gold-medal threshold. We release the two post-trained checkpoints as well as the training data, the training and inference code, the submitted solutions, and Nemotron-IMO-Bench, a new benchmark of 200 novel olympiad-level problems.",
                "url": "https://arxiv.org/abs/2609.10712",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        },
        {
            "id": "c4c2be61585c1148b846eb557a4b4368ca055646",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "Finishing the Task Is Not Enough: Evaluating Agent Resilience and Considerate Participation under Accumulating Challenge",
            "summary": "arXiv:2609.10724v1 Announce Type: new Abstract: Sustained deployment of generative AI agents requires more than isolated task success. Agents must remain useful across repeated interactions, changing conditions, and dependencies on people within shared workflows, especially as technical, human, and operational disruptions accumulate over time. We propose operational resilience and considerate participation as two complementary aspects of evaluating such agents: the former captures how agents recover from blocked work while preserving progress and communicating their limits, and the latter captures how their adaptation accounts for affected people, role boundaries, and the surrounding workflow. Yet both remain underexplored under accumulating challenge. We study 120 simulated healthcare trajectories across two generative AI models and twelve stakeholder-derived tasks under light, medium,…",
            "url": "https://arxiv.org/abs/2609.10724",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "Finishing the Task Is Not Enough: Evaluating Agent Resilience and Considerate Participation under Accumulating Challenge",
                "summary": "arXiv:2609.10724v1 Announce Type: new \nAbstract: Sustained deployment of generative AI agents requires more than isolated task success. Agents must remain useful across repeated interactions, changing conditions, and dependencies on people within shared workflows, especially as technical, human, and operational disruptions accumulate over time. We propose operational resilience and considerate participation as two complementary aspects of evaluating such agents: the former captures how agents recover from blocked work while preserving progress and communicating their limits, and the latter captures how their adaptation accounts for affected people, role boundaries, and the surrounding workflow. Yet both remain underexplored under accumulating challenge. We study 120 simulated healthcare trajectories across two generative AI models and twelve stakeholder-derived tasks under light, medium, and heavy challenge. We compare textual action plans, prompted internal assessments, and quantitative structured workload and affect reports to examine how agent behavior and reported state change as challenge accumulates. Regarding operational resilience, agents shift from self-directed recovery toward greater human dependence, while reporting increasing workload and negative affect in structured reports but seldom expressing strain in textual responses. Regarding considerate participation, agents broaden from task-focused adaptation toward task reframing, attention to others, role-boundary adjustment, and wider coordination, with distinct patterns across actions and internal assessments. From these findings, we derive five deployment dilemmas involving persistence, attention, role boundaries, state disclosure, and escalation that require stakeholder specification, further informing technical implications for learning, situated evaluation, and embodied adaptation.",
                "url": "https://arxiv.org/abs/2609.10724",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        },
        {
            "id": "c542f42f8638543912d723d2c1bb8a2531994148",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "Towards a Deterministic Math Solver for Clinical Language Models",
            "summary": "arXiv:2609.10728v1 Announce Type: new Abstract: Large language models are unreliable at arithmetic, which is a problem for clinical calculators where a single numerical error changes the recommendation. The standard response is to hardcode each calculator as a validated function, one at a time. We test an alternative: the model does not calculate. Instead, it writes case-specific Python that a restricted local executor runs as a deterministic solver, and the model's task reduces to deciding how to use it. We evaluate this Program-Solve interface on MedCalc-Bench Verified (1,100 cases, 55 calculators) against direct model arithmetic and a hand-written 22-calculator library, using Qwen2.5-7B and Qwen2.5-32B-AWQ, after auditing the benchmark's formulas against current clinical guidelines and flagging 16 of 55 with version, use or coefficient concerns. With formulas and gold variables suppl…",
            "url": "https://arxiv.org/abs/2609.10728",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "Towards a Deterministic Math Solver for Clinical Language Models",
                "summary": "arXiv:2609.10728v1 Announce Type: new \nAbstract: Large language models are unreliable at arithmetic, which is a problem for clinical calculators where a single numerical error changes the recommendation. The standard response is to hardcode each calculator as a validated function, one at a time. We test an alternative: the model does not calculate. Instead, it writes case-specific Python that a restricted local executor runs as a deterministic solver, and the model's task reduces to deciding how to use it. We evaluate this Program-Solve interface on MedCalc-Bench Verified (1,100 cases, 55 calculators) against direct model arithmetic and a hand-written 22-calculator library, using Qwen2.5-7B and Qwen2.5-32B-AWQ, after auditing the benchmark's formulas against current clinical guidelines and flagging 16 of 55 with version, use or coefficient concerns. With formulas and gold variables supplied and both routes reading the whole note, handing off to the solver is not a reliable advantage at 7B (75.31% against 72.02%, a paired +3.29 points with a 95% calculator-cluster interval of [-3.49, 10.38]) but is one at 32B (90.53% against 83.47%, +7.05 [0.47, 14.60], clear of zero). The hand-written library is exact on its 440 supported cases but abstains elsewhere (40.0% overall). Adding an executor thus helps some open-weight models more than others even under matched formula, variable and note access, and is not a substitute for verified formulas or reliable variable extraction either way.",
                "url": "https://arxiv.org/abs/2609.10728",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        },
        {
            "id": "371f68ee7f534511684d1c2b6d0ca1196e9ac342",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "Studying Without a Syllabus: Task-Agnostic Environment Preprocessing",
            "summary": "arXiv:2609.10824v1 Announce Type: new Abstract: Before an LLM agent tackles tasks in a new environment, it can inspect available corpora and tools and construct reusable resources such as indices, scripts, or procedural guidance. Most automated adaptation methods, however, rely on task examples, trajectories, or evaluation feedback to decide what to build. Existing task-agnostic approaches avoid this supervision but commit in advance to a preparation strategy for a particular type of environment. We study a more open-ended setting: can an agent study an unfamiliar environment without a syllabus, i.e. before test time and without knowledge of the downstream task distribution, and choose how to prepare it? We formalize task-agnostic environment preprocessing, in which a studying system explores an environment under a budget and produces artifacts for a frozen solver. We compare unaided an…",
            "url": "https://arxiv.org/abs/2609.10824",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "Studying Without a Syllabus: Task-Agnostic Environment Preprocessing",
                "summary": "arXiv:2609.10824v1 Announce Type: new \nAbstract: Before an LLM agent tackles tasks in a new environment, it can inspect available corpora and tools and construct reusable resources such as indices, scripts, or procedural guidance. Most automated adaptation methods, however, rely on task examples, trajectories, or evaluation feedback to decide what to build. Existing task-agnostic approaches avoid this supervision but commit in advance to a preparation strategy for a particular type of environment. We study a more open-ended setting: can an agent study an unfamiliar environment without a syllabus, i.e. before test time and without knowledge of the downstream task distribution, and choose how to prepare it? We formalize task-agnostic environment preprocessing, in which a studying system explores an environment under a budget and produces artifacts for a frozen solver. We compare unaided and archive-equipped meta-agents with fixed synthetic-practice and corpus-processing methods across six heterogeneous benchmarks. A meta-agent variant achieves the highest Avg@3 reward on five benchmarks, while fixed corpus processing remains best on the largest corpus benchmark. Larger study budgets do not reliably improve downstream reward. Nevertheless, studied artifacts reduce the test-time sampling needed to reach a given score, demonstrating how reusable preparation can shift computation from repeated test-time attempts to a pre-task study phase.",
                "url": "https://arxiv.org/abs/2609.10824",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        },
        {
            "id": "ab60aa784b8b8c9ed90c0d65109a64f9886fe16a",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "When Validation Stops Learning: Auditing Update Admission for Continual Embodied Agents",
            "summary": "arXiv:2609.10873v1 Announce Type: new Abstract: Independent evaluation can reject harmful policy updates yet also prevent useful continual learning. We argue that update admission must be assessed through both error control and retained learning opportunities at a stated interaction budget. We identify a concrete failure: a range-based confidence gate cannot certify unchanged old-task behavior within otherwise substantial budgets. A standard paired-binomial construction reduces this burden when outcome disagreements are rare. We also specify certified historical-reference promotion and a round-level missed-opportunity metric. In a constructed one-step pushing diagnostic with 32 seeds, fresh paired checks admit 31.6% of a common update stream at 2,000 episodes per stage, versus zero for the range-based gate; unconditional replay nevertheless learns better in closed-loop runs. A separate …",
            "url": "https://arxiv.org/abs/2609.10873",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "When Validation Stops Learning: Auditing Update Admission for Continual Embodied Agents",
                "summary": "arXiv:2609.10873v1 Announce Type: new \nAbstract: Independent evaluation can reject harmful policy updates yet also prevent useful continual learning. We argue that update admission must be assessed through both error control and retained learning opportunities at a stated interaction budget. We identify a concrete failure: a range-based confidence gate cannot certify unchanged old-task behavior within otherwise substantial budgets. A standard paired-binomial construction reduces this burden when outcome disagreements are rare. We also specify certified historical-reference promotion and a round-level missed-opportunity metric. In a constructed one-step pushing diagnostic with 32 seeds, fresh paired checks admit 31.6% of a common update stream at 2,000 episodes per stage, versus zero for the range-based gate; unconditional replay nevertheless learns better in closed-loop runs. A separate learned-dynamics stress test distinguishes model bias from feedback-selection error. The contribution is an admission-audit protocol with analytical and synthetic evidence; physical-robot and VLA validation remain open.",
                "url": "https://arxiv.org/abs/2609.10873",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        },
        {
            "id": "2f2d952764868736b0bde63273f1c2c79350f449",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "Decoupling Readiness from Release for Tail-Aware Scheduling of Agentic LLM Workflows",
            "summary": "arXiv:2609.10964v1 Announce Type: new Abstract: Agentic LLM workflows consist of sequences of model turns interleaved with tool interactions, so their end-to-end completion time depends not only on inference speed but also on when ready turns are released. Most runtimes release each turn immediately upon readiness. Under contention, this eager release policy can accumulate released but unfinished work; once submitted, those turns can no longer be reordered by the workflow-level policy, increasing tail latency. We present a tail-risk-aware turn release scheduling method that jointly decides which ready turn to release next and how much released but unfinished work to maintain. The method uses a mean--Conditional Value-at-Risk (CVaR) objective to capture the evolving tail risk of unfinished workflows, incorporates online estimates of turn work when prioritizing ready turns, and adapts the…",
            "url": "https://arxiv.org/abs/2609.10964",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "Decoupling Readiness from Release for Tail-Aware Scheduling of Agentic LLM Workflows",
                "summary": "arXiv:2609.10964v1 Announce Type: new \nAbstract: Agentic LLM workflows consist of sequences of model turns interleaved with tool interactions, so their end-to-end completion time depends not only on inference speed but also on when ready turns are released. Most runtimes release each turn immediately upon readiness. Under contention, this eager release policy can accumulate released but unfinished work; once submitted, those turns can no longer be reordered by the workflow-level policy, increasing tail latency. We present a tail-risk-aware turn release scheduling method that jointly decides which ready turn to release next and how much released but unfinished work to maintain. The method uses a mean--Conditional Value-at-Risk (CVaR) objective to capture the evolving tail risk of unfinished workflows, incorporates online estimates of turn work when prioritizing ready turns, and adapts the released work budget to observed queue pressure. We evaluate the method using real agent execution traces from software engineering tasks across multiple LLMs and workflow arrival rates. The method performs comparably to eager release under light load and substantially reduces the P95 of workflow flow time under contention, achieving up to a \\(3.50\\times\\) speedup.",
                "url": "https://arxiv.org/abs/2609.10964",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        },
        {
            "id": "4370e37fa4d5e35a3e7f060f361ebfd9d649fb5a",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "Demystifying the Privacy-Utility Trade-off in LLM Interactions",
            "summary": "arXiv:2609.10992v1 Announce Type: new Abstract: The integration of Large Language Models into daily tasks relies on context-rich instructions, inevitably exposing sensitive user information. Current privacy-preserving methods typically employ context-agnostic static rules, causing severe utility degradation. However, the specific mechanisms governing how sanitization impacts downstream performance remain largely underexplored. To address this, we conduct a systematic analysis to deconstruct the privacy-utility trade-off, uncovering three underlying mechanisms: (1) Context-Dependent Utility, which first establishes when to sanitize by revealing that data value shifts from critical constraints to dispensable noise based on user intent; (2) Strategic Adaptation, which subsequently determines how to sanitize by dictating that the choice between removal and replacement depends on the task's …",
            "url": "https://arxiv.org/abs/2609.10992",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "Demystifying the Privacy-Utility Trade-off in LLM Interactions",
                "summary": "arXiv:2609.10992v1 Announce Type: new \nAbstract: The integration of Large Language Models into daily tasks relies on context-rich instructions, inevitably exposing sensitive user information. Current privacy-preserving methods typically employ context-agnostic static rules, causing severe utility degradation. However, the specific mechanisms governing how sanitization impacts downstream performance remain largely underexplored. To address this, we conduct a systematic analysis to deconstruct the privacy-utility trade-off, uncovering three underlying mechanisms: (1) Context-Dependent Utility, which first establishes when to sanitize by revealing that data value shifts from critical constraints to dispensable noise based on user intent; (2) Strategic Adaptation, which subsequently determines how to sanitize by dictating that the choice between removal and replacement depends on the task's reliance on factual integrity versus structural coherence; and (3) Combinatorial Interplay, which finally extends the protection scope by demonstrating that attributes form a semantic web of synergistic dependencies or antagonistic redundancies. Guided by these insights, we introduce an intent-driven local protection framework. By distilling a lightweight model Veilmind-4B to drive a dynamic extraction-sanitization-restoration pipeline, our approach reaches a low-leakage privacy point while preserving substantially higher response utility than existing privacy-oriented baselines, advancing the privacy-utility trade-off toward the Pareto frontier.",
                "url": "https://arxiv.org/abs/2609.10992",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        },
        {
            "id": "52c24b56dba742c8822196696b5ab5b7ec4fac24",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "Defining AI Agents: A Compendium of Criteria, Metrics, and Benchmarks",
            "summary": "arXiv:2609.11018v1 Announce Type: new Abstract: The term agent in artificial intelligence lacks a standard definition, complicating the evaluation, comparison, and reproducibility of AI agent research. We address this ambiguity through a survey organized around five dimensions of agenticness: environmental interaction, learning and adaptation, autonomy, goal-directed behavior, and temporal coherence. For each dimension, we examine how the underlying capability has been conceptualized across prior work and synthesize the metrics, benchmarks, and evaluation frameworks used to assess it. This review provides a structured account of the current landscape of agent evaluation, highlighting both established approaches and areas where evaluation remains limited or inconsistent. We additionally introduce the Agent Compendium, a public-facing digital resource that organizes and extends the evalua…",
            "url": "https://arxiv.org/abs/2609.11018",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "Defining AI Agents: A Compendium of Criteria, Metrics, and Benchmarks",
                "summary": "arXiv:2609.11018v1 Announce Type: new \nAbstract: The term agent in artificial intelligence lacks a standard definition, complicating the evaluation, comparison, and reproducibility of AI agent research. We address this ambiguity through a survey organized around five dimensions of agenticness: environmental interaction, learning and adaptation, autonomy, goal-directed behavior, and temporal coherence. For each dimension, we examine how the underlying capability has been conceptualized across prior work and synthesize the metrics, benchmarks, and evaluation frameworks used to assess it. This review provides a structured account of the current landscape of agent evaluation, highlighting both established approaches and areas where evaluation remains limited or inconsistent. We additionally introduce the Agent Compendium, a public-facing digital resource that organizes and extends the evaluation methods identified through this review. Together, the survey and compendium provide a common structure for evaluating and comparing agent capabilities across AI systems, supporting more reproducible research, clearer communication, and more systematic study of artificial agents.",
                "url": "https://arxiv.org/abs/2609.11018",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        },
        {
            "id": "c7a8a3f60f07aac3c0b96a91ac5cce0857ee6b2e",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "The Agent Incident Registry: Toward Preventing Repeated AI Agent Failures",
            "summary": "arXiv:2609.11030v1 Announce Type: new Abstract: AI agents increasingly act through tools and delegated authority, but general incident repositories rarely capture the mechanisms needed to compare public failures with agent-security evaluations. We present the Agent Incident Registry (AIR), a source-linked catalog containing \\N{} records of agent-related events disclosed from \\Yfirst{} through \\Ylast{}. Each record includes supporting evidence, a stable identifier, and missingness-aware labels for causal role, disclosure class, mechanism, and outcome. Among the \\Nprimary{} generative-system records in which the agent acted, \\Rprimary{} involved realized harm (\\Pprimary\\%). Realized outcomes concentrate in in-the-wild and safety-failure records, while responsible disclosures and research demonstrations are overwhelmingly demonstrated; the aggregate share therefore characterizes collection…",
            "url": "https://arxiv.org/abs/2609.11030",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "The Agent Incident Registry: Toward Preventing Repeated AI Agent Failures",
                "summary": "arXiv:2609.11030v1 Announce Type: new \nAbstract: AI agents increasingly act through tools and delegated authority, but general incident repositories rarely capture the mechanisms needed to compare public failures with agent-security evaluations. We present the Agent Incident Registry (AIR), a source-linked catalog containing \\N{} records of agent-related events disclosed from \\Yfirst{} through \\Ylast{}. Each record includes supporting evidence, a stable identifier, and missingness-aware labels for causal role, disclosure class, mechanism, and outcome. Among the \\Nprimary{} generative-system records in which the agent acted, \\Rprimary{} involved realized harm (\\Pprimary\\%). Realized outcomes concentrate in in-the-wild and safety-failure records, while responsible disclosures and research demonstrations are overwhelmingly demonstrated; the aggregate share therefore characterizes collection composition rather than deployment risk. After initial curation, a second human reviewer checked all \\N{} records and their existing labels for completeness and correctness. In a deployment-analogue audit, InjecAgent's \\NInjecAgentCases{} cases occupy three of AIR's twelve surfaces and are all attacker-triggered, whereas AIR contains \\Nsafety{} no-adversary safety failures. AIR supports source-grounded case retrieval and evaluation-scope auditing, not failure-rate or control-efficacy estimation.",
                "url": "https://arxiv.org/abs/2609.11030",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        },
        {
            "id": "29d5c846f8310bc25b97fac015517e7ff5c33fa4",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "Grounding Agent Memory: Environment-Probing Curation for Enterprise Agents",
            "summary": "arXiv:2609.11060v1 Announce Type: new Abstract: Persistent memory is entering production-oriented agent platforms to help long-horizon agents accumulate experience across sessions. Yet a post-task curator agent restricted to completed trajectories can preserve errors, overgeneralize partial evidence, or retain stale knowledge. We introduce environment-probing curation, a deployment-compatible extension that gives an existing asynchronous curator agent least-privilege, read-only world tools to check, scope, and refresh candidate memories. It requires no model retraining and leaves the task agent, retriever, memory representation, and production write authority unchanged. In a production-like GitHub Copilot (GHCP) harness built on its SDK, we compare stateless execution, full in-context learning, GHCP + Mem, and GHCP + Mem (w/ Env Probing) on CLBench database exploration and 90 adapted AP…",
            "url": "https://arxiv.org/abs/2609.11060",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "Grounding Agent Memory: Environment-Probing Curation for Enterprise Agents",
                "summary": "arXiv:2609.11060v1 Announce Type: new \nAbstract: Persistent memory is entering production-oriented agent platforms to help long-horizon agents accumulate experience across sessions. Yet a post-task curator agent restricted to completed trajectories can preserve errors, overgeneralize partial evidence, or retain stale knowledge. We introduce environment-probing curation, a deployment-compatible extension that gives an existing asynchronous curator agent least-privilege, read-only world tools to check, scope, and refresh candidate memories. It requires no model retraining and leaves the task agent, retriever, memory representation, and production write authority unchanged. In a production-like GitHub Copilot (GHCP) harness built on its SDK, we compare stateless execution, full in-context learning, GHCP + Mem, and GHCP + Mem (w/ Env Probing) on CLBench database exploration and 90 adapted APEX management-consulting tasks. On CLBench, probing raises pass rate from 39% to 73% and pass-discounted reward from 8.60 to 22.60 while reducing queries from 8.8 to 4.7 per question and task-agent cost from \\$3.38 to \\$1.68. Across six APEX worlds, all 18 memory-versus-baseline mean reward comparisons are positive and task-agent tool calls fall by 16--75%; probing gives the best task-agent reward gain per dollar in five worlds. Probing also attains higher mean reward than GHCP + Mem on both Sonnet 4.6 and Opus 4.7 without schema drift. Environment probing therefore turns existing agent-memory curation into an environment-informed, auditable process while preserving a compact task-time interface.",
                "url": "https://arxiv.org/abs/2609.11060",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        },
        {
            "id": "894fbeb62b581d2b8a75e5085f02b99207ba77a9",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "Fork Where the Model Changes Its Mind: Belief-Shift Branching for Tree-Structured Reinforcement Learning",
            "summary": "arXiv:2609.11061v1 Announce Type: new Abstract: Tree-structured rollouts give critic-free reinforcement learning with verifiable rewards (RLVR) step-level credit: fork a chain at an intermediate point, and sibling outcome differences estimate step value. Each fork adds sampling cost, so realistic budgets typically allow only a few forks per chain. A fork placed where the outcome is already largely settled yields siblings that mostly agree and provide almost no credit signal; hence, for a given tree size, where forks are placed largely determines how much step-level RL can gain. Most existing mainstream methods place forks by structure, such as fixed lengths, midpoints, and delimiters, or by next-token entropy. We formalize fork placement as locating the \\emph{pivots} of the chain's value curve, where the expected outcome turns. We propose \\emph{belief-shift branching}: read the model's …",
            "url": "https://arxiv.org/abs/2609.11061",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "Fork Where the Model Changes Its Mind: Belief-Shift Branching for Tree-Structured Reinforcement Learning",
                "summary": "arXiv:2609.11061v1 Announce Type: new \nAbstract: Tree-structured rollouts give critic-free reinforcement learning with verifiable rewards (RLVR) step-level credit: fork a chain at an intermediate point, and sibling outcome differences estimate step value. Each fork adds sampling cost, so realistic budgets typically allow only a few forks per chain. A fork placed where the outcome is already largely settled yields siblings that mostly agree and provide almost no credit signal; hence, for a given tree size, where forks are placed largely determines how much step-level RL can gain. Most existing mainstream methods place forks by structure, such as fixed lengths, midpoints, and delimiters, or by next-token entropy. We formalize fork placement as locating the \\emph{pivots} of the chain's value curve, where the expected outcome turns. We propose \\emph{belief-shift branching}: read the model's answer belief at candidate boundaries and fork just before the step where consecutive beliefs diverge most. Three instantiations, none needing step-level supervision, span access levels: a black-box probe, a logit-lens depth profile, and a learned activation direction, which is fit offline and therefore used only in the validation before RL training. The signal only \\emph{places} forks, and the probe costs about $1\\%$ of step compute on mathematics and under $5\\%$ on code when it runs inside the rollout engine. In that validation, against Monte-Carlo value curves, a belief-shift signal ranks first in each of the eight model$\\times$benchmark panels, ahead of entropy, structural, and LLM-judge baselines. In RL across three model families and two domains, belief-shift forking leads every mathematics aggregate, on OLMo-3-7B by $+2.6$ aggregate and $+2.9$ on AIME 2026 over the strongest baseline, and sweeps every OLMo code column, by $+6.5$ on LiveCodeBench-medium.",
                "url": "https://arxiv.org/abs/2609.11061",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        },
        {
            "id": "78f499d5113d760585d6c9263e0102d89599b8ec",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "MOSAIC: Query-Aware Exploration Policy Adaptation for GraphRAG",
            "summary": "arXiv:2609.11065v1 Announce Type: new Abstract: Graph Retrieval-Augmented Generation (GraphRAG) can connect evidence distributed across a corpus graph, but most systems use largely shared exploration procedures across queries. This creates a structural mismatch: direct facts may need compact local neighborhoods, comparisons need balanced coverage of multiple targets, and mediated questions may require deeper paths through weakly related connectors. We present Mosaic, a training-free framework that formulates GraphRAG retrieval as a per-query control problem. An LLM analyzer converts query-specific evidence requirements into a bounded policy over seed selection, graph traversal, stopping, and evidence selection, while the corpus graph, indexes, scoring functions, grounding procedure, and answer generator remain shared. On GraphRAG-Bench, Mosaic achieves query-weighted Answer Correctness …",
            "url": "https://arxiv.org/abs/2609.11065",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "MOSAIC: Query-Aware Exploration Policy Adaptation for GraphRAG",
                "summary": "arXiv:2609.11065v1 Announce Type: new \nAbstract: Graph Retrieval-Augmented Generation (GraphRAG) can connect evidence distributed across a corpus graph, but most systems use largely shared exploration procedures across queries. This creates a structural mismatch: direct facts may need compact local neighborhoods, comparisons need balanced coverage of multiple targets, and mediated questions may require deeper paths through weakly related connectors. We present Mosaic, a training-free framework that formulates GraphRAG retrieval as a per-query control problem. An LLM analyzer converts query-specific evidence requirements into a bounded policy over seed selection, graph traversal, stopping, and evidence selection, while the corpus graph, indexes, scoring functions, grounding procedure, and answer generator remain shared.\n  On GraphRAG-Bench, Mosaic achieves query-weighted Answer Correctness of 76.97 on Medical and 64.33 on Novel, improving over the strongest previously reported overall results by 5.13 and 4.43 points. On Medical, it reaches 95.1 Evidence Recall and 86.1 Context Relevancy. Controlled comparisons on an identical graph and generator show that no fixed narrow, medium, or wide policy is consistently optimal; Mosaic improves by 9.96 points over the strongest canonical fixed policy. Relative to Fixed Wide, it evaluates 81.9% fewer paths and retains 47.2% fewer evidence items. Transfer experiments on HotpotQA, MuSiQue, and 2WikiMultiHopQA further show that the policy interface can be applied without benchmark-specific retriever training.",
                "url": "https://arxiv.org/abs/2609.11065",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        },
        {
            "id": "98937f50b79f911ed6bd126075646b290ad83072",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "Benchmark Radar: A Living Database and Search Engine for AI Benchmarks and Evaluation",
            "summary": "arXiv:2609.11115v1 Announce Type: new Abstract: Benchmark researchers and developers of large language models (LLMs) and other AI systems need to find relevant evaluations, locate their benchmark datasets and code, and understand the settings behind reported scores. We present Benchmark Radar, a living database and search engine for retrieval and discovery of AI benchmarks, covering LLM evaluation, agentic and tool-use benchmarks, coding, reasoning, safety, and domain-specific evaluations. The system combines daily discovery of benchmark papers, repositories, datasets, and releases with a searchable benchmark catalog, mentions in model cards and technical reports, and score histories. It retains source identities and citations so readers can inspect candidate benchmarks and their evaluation evidence. Daily discovery draws on 37 sources: 13 direct connectors and 24 first-party research a…",
            "url": "https://arxiv.org/abs/2609.11115",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "Benchmark Radar: A Living Database and Search Engine for AI Benchmarks and Evaluation",
                "summary": "arXiv:2609.11115v1 Announce Type: new \nAbstract: Benchmark researchers and developers of large language models (LLMs) and other AI systems need to find relevant evaluations, locate their benchmark datasets and code, and understand the settings behind reported scores. We present Benchmark Radar, a living database and search engine for retrieval and discovery of AI benchmarks, covering LLM evaluation, agentic and tool-use benchmarks, coding, reasoning, safety, and domain-specific evaluations. The system combines daily discovery of benchmark papers, repositories, datasets, and releases with a searchable benchmark catalog, mentions in model cards and technical reports, and score histories. It retains source identities and citations so readers can inspect candidate benchmarks and their evaluation evidence. Daily discovery draws on 37 sources: 13 direct connectors and 24 first-party research and engineering feeds. The catalog contains 1,283 source records drawn from 4 benchmark catalogs and 12,916 numeric observations on 790 records. We describe collection and retrieval, audit the full catalog, and examine benchmark saturation, adoption trends, and the limits of score comparisons. A worked example walks through a complete prior-art search, showing how to query the catalog and inspect benchmark evidence when designing a new evaluation. We release the web dashboard with a benchmark leaderboard, a Pareto frontier view of score against measured use, saturation and trend views, daily feeds, downloadable evidence, a command-line interface (CLI) for offline queries, and reproducible analysis.",
                "url": "https://arxiv.org/abs/2609.11115",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        },
        {
            "id": "a1e8ba4e259ad54495c8104d1185eb6ff8eead28",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "KuaiRP Series Role-playing Models Technical Report",
            "summary": "arXiv:2609.11127v1 Announce Type: new Abstract: This paper introduces the complete technical solution for the KuaiRP series of role-playing models. We aim to achieve four core objectives for a dedicated role-playing model: simplified prompt engineering, highly stable output quality, built-in domain world knowledge, and high-efficiency deployment with a small parameter size. However, effectively injecting deep domain knowledge often leads to a severe catastrophic forgetting of the model's general agent capabilities. To overcome this trade-off, we propose a multi-stage training pipeline. First, we design a standardized character template and construct an SFT data pipeline based on user behavior simulation and reverse profile filtering. Next, we utilize a rule-based composite reward function during the Reinforcement Learning (RL) phase to eliminate common degradation phenomena like length …",
            "url": "https://arxiv.org/abs/2609.11127",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "KuaiRP Series Role-playing Models Technical Report",
                "summary": "arXiv:2609.11127v1 Announce Type: new \nAbstract: This paper introduces the complete technical solution for the KuaiRP series of role-playing models. We aim to achieve four core objectives for a dedicated role-playing model: simplified prompt engineering, highly stable output quality, built-in domain world knowledge, and high-efficiency deployment with a small parameter size. However, effectively injecting deep domain knowledge often leads to a severe catastrophic forgetting of the model's general agent capabilities. To overcome this trade-off, we propose a multi-stage training pipeline. First, we design a standardized character template and construct an SFT data pipeline based on user behavior simulation and reverse profile filtering. Next, we utilize a rule-based composite reward function during the Reinforcement Learning (RL) phase to eliminate common degradation phenomena like length expansion and repetitive generation. Finally, to recover the general capabilities compromised during SFT and RL, we propose a novel self-distillation paradigm using Two-stage On-Policy Distillation (OPD) equipped with Cumulative-Divergence Decay (CDD). By using the domain-adapted model as the teacher and the original base model as the student, we effectively balance deep domain knowledge injection with the preservation of general agent capabilities. Experimental results demonstrate that the KuaiRP models not only match the current state-of-the-art proprietary models in role-playing fidelity within our target domains, but also successfully recover general agent capabilities, maintaining extremely low deployment costs.",
                "url": "https://arxiv.org/abs/2609.11127",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        },
        {
            "id": "4b2a6ee44d7402fdcee9e161138b5ddb2c4b2496",
            "source_id": "arxiv_ai",
            "source": "arXiv — Artificial Intelligence",
            "category": "ai",
            "engine": "rss",
            "title": "Same Day, Same Story; One Day Ahead, a Different Signal: The Dual Validity of Financial Sentiment",
            "summary": "arXiv:2609.11144v1 Announce Type: new Abstract: Financial NLP has a standard workflow: validate a sentiment tool against human labels, then trust it to extract market signal. This assumes the two evaluations measure the same thing. We test that assumption in a setting where both can be measured at once: a corpus of securities class actions (2002-2025) linking 70,500 X messages to abnormal stock returns, with a single-annotator human labelled gold sample. Running five instruments (VADER, Loughran-McDonald, FinBERT, Twitter-RoBERTa, and an LLM annotator) through one identical pipeline, we find that the relationship between construct and predictive validity depends on the sampling convention and score representation. Under conventional method-specific sampling, human agreement aligns more closely with graded same-day associations than with one-day leads. On a fixed-n panel, however, agreem…",
            "url": "https://arxiv.org/abs/2609.11144",
            "image": "",
            "published": "2026-09-12T04:00:00+00:00",
            "score": 78.96,
            "color": "#b09cff",
            "raw": {
                "title": "Same Day, Same Story; One Day Ahead, a Different Signal: The Dual Validity of Financial Sentiment",
                "summary": "arXiv:2609.11144v1 Announce Type: new \nAbstract: Financial NLP has a standard workflow: validate a sentiment tool against human labels, then trust it to extract market signal. This assumes the two evaluations measure the same thing. We test that assumption in a setting where both can be measured at once: a corpus of securities class actions (2002-2025) linking 70,500 X messages to abnormal stock returns, with a single-annotator human labelled gold sample. Running five instruments (VADER, Loughran-McDonald, FinBERT, Twitter-RoBERTa, and an LLM annotator) through one identical pipeline, we find that the relationship between construct and predictive validity depends on the sampling convention and score representation. Under conventional method-specific sampling, human agreement aligns more closely with graded same-day associations than with one-day leads. On a fixed-n panel, however, agreement has similar graded rank correlations at both horizons, while the coarse ordering remains weak. Benchmark agreement therefore establishes semantic validity but does not by itself determine predictive rankings. In a conversation that is 17.6% spam, message volume predicts neither market damage nor settlement size.",
                "url": "https://arxiv.org/abs/2609.11144",
                "image": "",
                "published": "Sat, 12 Sep 2026 00:00:00 -0400"
            }
        }
    ],
    "stored": 0,
    "ai": {
        "enriched": false,
        "stored": 0
    },
    "health": [
        {
            "id": "arxiv_ai",
            "name": "arXiv — Artificial Intelligence",
            "engine": "rss",
            "category": "ai",
            "ok": true,
            "error": "",
            "ms": 125,
            "count": 20,
            "cached_at": "2026-09-12T20:32:46+00:00"
        }
    ]
}