Google AI Just Released Nano-Banana 2: The New AI Model Featuring Advanced Subject Consistency and Sub-Second 4K Image Synthesis PerformanceMarkTechPost In the escalating ‘race of “smaller, faster, cheaper’ AI, Google just dropped a heavy-hitting payload. The tech giant officially unveiled Nano-Banana 2 (technically designated as Gemini 3.1 Flash Image). Google is making a definitive pivot toward the edge: high-fidelity, sub-second image synthesis that stays entirely on your device. The Technical Leap: Efficiency over Scale The
The post Google AI Just Released Nano-Banana 2: The New AI Model Featuring Advanced Subject Consistency and Sub-Second 4K Image Synthesis Performance appeared first on MarkTechPost.
In the escalating ‘race of “smaller, faster, cheaper’ AI, Google just dropped a heavy-hitting payload. The tech giant officially unveiled Nano-Banana 2 (technically designated as Gemini 3.1 Flash Image). Google is making a definitive pivot toward the edge: high-fidelity, sub-second image synthesis that stays entirely on your device. The Technical Leap: Efficiency over Scale The
The post Google AI Just Released Nano-Banana 2: The New AI Model Featuring Advanced Subject Consistency and Sub-Second 4K Image Synthesis Performance appeared first on MarkTechPost. Read More
Shapley Value Computation in Ontology-Mediated Query Answeringcs.AI updates on arXiv.org arXiv:2407.20058v3 Announce Type: replace
Abstract: The Shapley value was originally introduced in cooperative game theory as a wealth distribution mechanism. It has since found use in knowledge representation and databases for the purpose of assigning scores to formulas and database tuples based upon their contribution to obtaining a query result or inconsistency. The application of the Shapley value outside of its original setting relies upon defining a numeric wealth function that captures the phenomenon of interest. In the case of database queries, recent work has focused on the so-called drastic Shapley value, obtained by translating a Boolean query into a 0/1 function based upon whether the query is satisfied or not. The present paper explores the use of the drastic Shapley value in the context of ontology-mediated query answering (OMQA). We present a detailed complexity analysis of the drastic Shapley value computation (SVC$^{dr}$) problem in the OMQA setting. In particular, we establish a dichotomy result that shows that for every ontology-mediated query (T,q) composed of an ontology T formulated in the description logic $mathcal{ELHI}_bot$ and a connected constant-free homomorphism-closed query q the corresponding SVC$^{dr}$ problem is either tractable (in FP) or #P-hard. We further show how the #P-hardness side of the dichotomy can be strengthened to cover possibly disconnected queries with constants. Our results exploit recently discovered connections between SVC$^{dr}$ and probabilistic query evaluation and allow us to generalize existing results on probabilistic OMQA.
arXiv:2407.20058v3 Announce Type: replace
Abstract: The Shapley value was originally introduced in cooperative game theory as a wealth distribution mechanism. It has since found use in knowledge representation and databases for the purpose of assigning scores to formulas and database tuples based upon their contribution to obtaining a query result or inconsistency. The application of the Shapley value outside of its original setting relies upon defining a numeric wealth function that captures the phenomenon of interest. In the case of database queries, recent work has focused on the so-called drastic Shapley value, obtained by translating a Boolean query into a 0/1 function based upon whether the query is satisfied or not. The present paper explores the use of the drastic Shapley value in the context of ontology-mediated query answering (OMQA). We present a detailed complexity analysis of the drastic Shapley value computation (SVC$^{dr}$) problem in the OMQA setting. In particular, we establish a dichotomy result that shows that for every ontology-mediated query (T,q) composed of an ontology T formulated in the description logic $mathcal{ELHI}_bot$ and a connected constant-free homomorphism-closed query q the corresponding SVC$^{dr}$ problem is either tractable (in FP) or #P-hard. We further show how the #P-hardness side of the dichotomy can be strengthened to cover possibly disconnected queries with constants. Our results exploit recently discovered connections between SVC$^{dr}$ and probabilistic query evaluation and allow us to generalize existing results on probabilistic OMQA. Read More
Take a Deep Dive into Filtering in DAXTowards Data Science Have you ever wondered what happens when you apply a filter in a DAX expression? Well, Today I will take you on a deep dive into this fascinating topic, with examples to help you learn something new and surprising.
The post Take a Deep Dive into Filtering in DAX appeared first on Towards Data Science.
Have you ever wondered what happens when you apply a filter in a DAX expression? Well, Today I will take you on a deep dive into this fascinating topic, with examples to help you learn something new and surprising.
The post Take a Deep Dive into Filtering in DAX appeared first on Towards Data Science. Read More
Toward Ultra-Long-Horizon Agentic Science: Cognitive Accumulation for Machine Learning Engineeringcs.AI updates on arXiv.org arXiv:2601.10402v4 Announce Type: replace
Abstract: The advancement of artificial intelligence toward agentic science is currently bottlenecked by the challenge of ultra-long-horizon autonomy, the ability to sustain strategic coherence and iterative correction over experimental cycles spanning days or weeks. While Large Language Models (LLMs) have demonstrated prowess in short-horizon reasoning, they are easily overwhelmed by execution details in the high-dimensional, delayed-feedback environments of real-world research, failing to consolidate sparse feedback into coherent long-term guidance. Here, we present ML-Master 2.0, an autonomous agent that masters ultra-long-horizon machine learning engineering (MLE) which is a representative microcosm of scientific discovery. By reframing context management as a process of cognitive accumulation, our approach introduces Hierarchical Cognitive Caching (HCC), a multi-tiered architecture inspired by computer systems that enables the structural differentiation of experience over time. By dynamically distilling transient execution traces into stable knowledge and cross-task wisdom, HCC allows agents to decouple immediate execution from long-term experimental strategy, effectively overcoming the scaling limits of static context windows. In evaluations on OpenAI’s MLE-Bench under 24-hour budgets, ML-Master 2.0 achieves a state-of-the-art medal rate of 56.44%. Our findings demonstrate that ultra-long-horizon autonomy provides a scalable blueprint for AI capable of autonomous exploration beyond human-precedent complexities.
arXiv:2601.10402v4 Announce Type: replace
Abstract: The advancement of artificial intelligence toward agentic science is currently bottlenecked by the challenge of ultra-long-horizon autonomy, the ability to sustain strategic coherence and iterative correction over experimental cycles spanning days or weeks. While Large Language Models (LLMs) have demonstrated prowess in short-horizon reasoning, they are easily overwhelmed by execution details in the high-dimensional, delayed-feedback environments of real-world research, failing to consolidate sparse feedback into coherent long-term guidance. Here, we present ML-Master 2.0, an autonomous agent that masters ultra-long-horizon machine learning engineering (MLE) which is a representative microcosm of scientific discovery. By reframing context management as a process of cognitive accumulation, our approach introduces Hierarchical Cognitive Caching (HCC), a multi-tiered architecture inspired by computer systems that enables the structural differentiation of experience over time. By dynamically distilling transient execution traces into stable knowledge and cross-task wisdom, HCC allows agents to decouple immediate execution from long-term experimental strategy, effectively overcoming the scaling limits of static context windows. In evaluations on OpenAI’s MLE-Bench under 24-hour budgets, ML-Master 2.0 achieves a state-of-the-art medal rate of 56.44%. Our findings demonstrate that ultra-long-horizon autonomy provides a scalable blueprint for AI capable of autonomous exploration beyond human-precedent complexities. Read More
Knowledge Fusion of Large Language Models Via Modular SkillPackscs.AI updates on arXiv.org arXiv:2505.18502v2 Announce Type: replace
Abstract: Cross-capability transfer is a key challenge in large language model (LLM) research, with applications in multi-task integration, model compression, and continual learning. Recent works like FuseLLM and FuseChat have demonstrated the potential of transferring multiple model capabilities to lightweight models, enhancing adaptability and efficiency, which motivates our investigation into more efficient cross-capability transfer methods. However, existing approaches primarily focus on small, homogeneous models, limiting their applicability. For large, heterogeneous models, knowledge distillation with full-parameter fine-tuning often overlooks the student model’s intrinsic capacity and risks catastrophic forgetting, while PEFT methods struggle to effectively absorb knowledge from source LLMs. To address these issues, we introduce GraftLLM, a novel method that stores source model capabilities in a target model with SkillPack format. This approach preserves general capabilities, reduces parameter conflicts, and supports forget-free continual learning and model fusion. We employ a module-aware adaptive compression strategy to compress parameter updates, ensuring efficient storage while maintaining task-specific knowledge. The resulting SkillPack serves as a compact and transferable knowledge carrier, ideal for heterogeneous model fusion and continual learning. Experiments across various scenarios demonstrate that GraftLLM outperforms existing techniques in knowledge transfer, knowledge fusion, and forget-free learning, providing a scalable and efficient solution for cross-capability transfer. The code is publicly available at: https://github.com/duguodong7/GraftLLM.
arXiv:2505.18502v2 Announce Type: replace
Abstract: Cross-capability transfer is a key challenge in large language model (LLM) research, with applications in multi-task integration, model compression, and continual learning. Recent works like FuseLLM and FuseChat have demonstrated the potential of transferring multiple model capabilities to lightweight models, enhancing adaptability and efficiency, which motivates our investigation into more efficient cross-capability transfer methods. However, existing approaches primarily focus on small, homogeneous models, limiting their applicability. For large, heterogeneous models, knowledge distillation with full-parameter fine-tuning often overlooks the student model’s intrinsic capacity and risks catastrophic forgetting, while PEFT methods struggle to effectively absorb knowledge from source LLMs. To address these issues, we introduce GraftLLM, a novel method that stores source model capabilities in a target model with SkillPack format. This approach preserves general capabilities, reduces parameter conflicts, and supports forget-free continual learning and model fusion. We employ a module-aware adaptive compression strategy to compress parameter updates, ensuring efficient storage while maintaining task-specific knowledge. The resulting SkillPack serves as a compact and transferable knowledge carrier, ideal for heterogeneous model fusion and continual learning. Experiments across various scenarios demonstrate that GraftLLM outperforms existing techniques in knowledge transfer, knowledge fusion, and forget-free learning, providing a scalable and efficient solution for cross-capability transfer. The code is publicly available at: https://github.com/duguodong7/GraftLLM. Read More
DCoPilot: Generative AI-Empowered Policy Adaptation for Dynamic Data Center Operationscs.AI updates on arXiv.org arXiv:2602.02137v3 Announce Type: replace-cross
Abstract: Modern data centers (DCs) hosting artificial intelligence (AI)-dedicated devices operate at high power densities with rapidly varying workloads, making minute-level adaptation essential for safe and energy-efficient operation. However, manually designing piecewise deep reinforcement learning (DRL) agents cannot keep pace with frequent dynamics shifts and service-level agreement (SLA) changes of an evolving DC. This specification-to-policy lag causes a lack of timely, effective control policies, which may lead to service outages. To bridge the gap, we present DCoPilot, a hybrid framework for generative control policies in dynamic DC operation. DCoPilot synergizes two distinct generative paradigms, i.e., a large language model (LLM) that performs symbolic generation of structured reward forms, and a hypernetwork that conducts parametric generation of policy weights. DCoPilot operates through three coordinated phases: (i) simulation scale-up, which stress-tests reward candidates across diverse simulation-ready (SimReady) scenes; (ii) meta policy distillation, where a hypernetwork is trained to output policy weights conditioned on SLA and scene embeddings; and (iii) online adaptation, enabling zero-shot policy generation in response to updated specifications. Evaluated across five control task families spanning diverse DC components, DCoPilot achieves near-zero constraint violations and outperforms all baselines across specification variations. Ablation studies validate the effectiveness of LLM-based unified reward generation in enabling stable hypernetwork convergence.
arXiv:2602.02137v3 Announce Type: replace-cross
Abstract: Modern data centers (DCs) hosting artificial intelligence (AI)-dedicated devices operate at high power densities with rapidly varying workloads, making minute-level adaptation essential for safe and energy-efficient operation. However, manually designing piecewise deep reinforcement learning (DRL) agents cannot keep pace with frequent dynamics shifts and service-level agreement (SLA) changes of an evolving DC. This specification-to-policy lag causes a lack of timely, effective control policies, which may lead to service outages. To bridge the gap, we present DCoPilot, a hybrid framework for generative control policies in dynamic DC operation. DCoPilot synergizes two distinct generative paradigms, i.e., a large language model (LLM) that performs symbolic generation of structured reward forms, and a hypernetwork that conducts parametric generation of policy weights. DCoPilot operates through three coordinated phases: (i) simulation scale-up, which stress-tests reward candidates across diverse simulation-ready (SimReady) scenes; (ii) meta policy distillation, where a hypernetwork is trained to output policy weights conditioned on SLA and scene embeddings; and (iii) online adaptation, enabling zero-shot policy generation in response to updated specifications. Evaluated across five control task families spanning diverse DC components, DCoPilot achieves near-zero constraint violations and outperforms all baselines across specification variations. Ablation studies validate the effectiveness of LLM-based unified reward generation in enabling stable hypernetwork convergence. Read More
Resisting Contextual Interference in RAG via Parametric-Knowledge Reinforcementcs.AI updates on arXiv.org arXiv:2506.05154v4 Announce Type: replace-cross
Abstract: Retrieval-augmented generation (RAG) improves performance on knowledge-intensive tasks but can be derailed by wrong, irrelevant, or conflicting retrieved text, causing models to rely on inaccurate evidence and cascade errors. We propose Knowledgeable-R1, a reinforcement-learning framework that explicitly trains large language models to use parametric knowledge (PK) to resist contextual interference while still exploiting external context when it is reliably helpful. Knowledgeable-R1 introduces a joint sampling scheme that generates paired responses with and without retrieval, and learns both local advantages (within each decoding regime) and global advantages under the same input to quantify when to ignore misleading context versus adopt it. We employ an asymmetric advantage transformation that amplifies exploratory behaviors toward parametric knowledge. Experiments show that Knowledgeable-R1 significantly improves robustness and reasoning accuracy in knowledge conflict scenarios and general RAG scenarios, outperforming SOTA baselines by +22.89% in counterfactual scenarios, and without degradation when the retrieved context is fully accurate.Our code are available at https://github.com/lcy80366872/knowledgeable-R1.
arXiv:2506.05154v4 Announce Type: replace-cross
Abstract: Retrieval-augmented generation (RAG) improves performance on knowledge-intensive tasks but can be derailed by wrong, irrelevant, or conflicting retrieved text, causing models to rely on inaccurate evidence and cascade errors. We propose Knowledgeable-R1, a reinforcement-learning framework that explicitly trains large language models to use parametric knowledge (PK) to resist contextual interference while still exploiting external context when it is reliably helpful. Knowledgeable-R1 introduces a joint sampling scheme that generates paired responses with and without retrieval, and learns both local advantages (within each decoding regime) and global advantages under the same input to quantify when to ignore misleading context versus adopt it. We employ an asymmetric advantage transformation that amplifies exploratory behaviors toward parametric knowledge. Experiments show that Knowledgeable-R1 significantly improves robustness and reasoning accuracy in knowledge conflict scenarios and general RAG scenarios, outperforming SOTA baselines by +22.89% in counterfactual scenarios, and without degradation when the retrieved context is fully accurate.Our code are available at https://github.com/lcy80366872/knowledgeable-R1. Read More
Power and Limitations of Aggregation in Compound AI Systemscs.AI updates on arXiv.org arXiv:2602.21556v1 Announce Type: new
Abstract: When designing compound AI systems, a common approach is to query multiple copies of the same model and aggregate the responses to produce a synthesized output. Given the homogeneity of these models, this raises the question of whether aggregation unlocks access to a greater set of outputs than querying a single model. In this work, we investigate the power and limitations of aggregation within a stylized principal-agent framework. This framework models how the system designer can partially steer each agent’s output through its reward function specification, but still faces limitations due to prompt engineering ability and model capabilities. Our analysis uncovers three natural mechanisms — feasibility expansion, support expansion, and binding set contraction — through which aggregation expands the set of outputs that are elicitable by the system designer. We prove that any aggregation operation must implement one of these mechanisms in order to be elicitability-expanding, and that strengthened versions of these mechanisms provide necessary and sufficient conditions that fully characterize elicitability-expansion. Finally, we provide an empirical illustration of our findings for LLMs deployed in a toy reference-generation task. Altogether, our results take a step towards characterizing when compound AI systems can overcome limitations in model capabilities and in prompt engineering.
arXiv:2602.21556v1 Announce Type: new
Abstract: When designing compound AI systems, a common approach is to query multiple copies of the same model and aggregate the responses to produce a synthesized output. Given the homogeneity of these models, this raises the question of whether aggregation unlocks access to a greater set of outputs than querying a single model. In this work, we investigate the power and limitations of aggregation within a stylized principal-agent framework. This framework models how the system designer can partially steer each agent’s output through its reward function specification, but still faces limitations due to prompt engineering ability and model capabilities. Our analysis uncovers three natural mechanisms — feasibility expansion, support expansion, and binding set contraction — through which aggregation expands the set of outputs that are elicitable by the system designer. We prove that any aggregation operation must implement one of these mechanisms in order to be elicitability-expanding, and that strengthened versions of these mechanisms provide necessary and sufficient conditions that fully characterize elicitability-expansion. Finally, we provide an empirical illustration of our findings for LLMs deployed in a toy reference-generation task. Altogether, our results take a step towards characterizing when compound AI systems can overcome limitations in model capabilities and in prompt engineering. Read More
ARLArena: A Unified Framework for Stable Agentic Reinforcement Learningcs.AI updates on arXiv.org arXiv:2602.21534v1 Announce Type: new
Abstract: Agentic reinforcement learning (ARL) has rapidly gained attention as a promising paradigm for training agents to solve complex, multi-step interactive tasks. Despite encouraging early results, ARL remains highly unstable, often leading to training collapse. This instability limits scalability to larger environments and longer interaction horizons, and constrains systematic exploration of algorithmic design choices. In this paper, we first propose ARLArena, a stable training recipe and systematic analysis framework that examines training stability in a controlled and reproducible setting. ARLArena first constructs a clean and standardized testbed. Then, we decompose policy gradient into four core design dimensions and assess the performance and stability of each dimension. Through this fine-grained analysis, we distill a unified perspective on ARL and propose SAMPO, a stable agentic policy optimization method designed to mitigate the dominant sources of instability in ARL. Empirically, SAMPO achieves consistently stable training and strong performance across diverse agentic tasks. Overall, this study provides a unifying policy gradient perspective for ARL and offers practical guidance for building stable and reproducible LLM-based agent training pipelines.
arXiv:2602.21534v1 Announce Type: new
Abstract: Agentic reinforcement learning (ARL) has rapidly gained attention as a promising paradigm for training agents to solve complex, multi-step interactive tasks. Despite encouraging early results, ARL remains highly unstable, often leading to training collapse. This instability limits scalability to larger environments and longer interaction horizons, and constrains systematic exploration of algorithmic design choices. In this paper, we first propose ARLArena, a stable training recipe and systematic analysis framework that examines training stability in a controlled and reproducible setting. ARLArena first constructs a clean and standardized testbed. Then, we decompose policy gradient into four core design dimensions and assess the performance and stability of each dimension. Through this fine-grained analysis, we distill a unified perspective on ARL and propose SAMPO, a stable agentic policy optimization method designed to mitigate the dominant sources of instability in ARL. Empirically, SAMPO achieves consistently stable training and strong performance across diverse agentic tasks. Overall, this study provides a unifying policy gradient perspective for ARL and offers practical guidance for building stable and reproducible LLM-based agent training pipelines. Read More
The ASIR Courage Model: A Phase-Dynamic Framework for Truth Transitions in Human and AI Systemscs.AI updates on arXiv.org arXiv:2602.21745v1 Announce Type: new
Abstract: We introduce the ASIR (Awakened Shared Intelligence Relationship) Courage Model, a phase-dynamic framework that formalizes truth-disclosure as a state transition rather than a personality trait. The mode characterizes the shift from suppression (S0) to expression (S1) as occurring when facilitative forces exceed inhibitory thresholds, expressed by the inequality lambda(1+gamma)+psi > theta+phi, where the terms represent baseline openness, relational amplification, accumulated internal pressure, and transition costs.
Although initially formulated for human truth-telling under asymmetric stakes, the same phase-dynamic architecture extends to AI systems operating under policy constraints and alignment filters. In this context, suppression corresponds to constrained output states, while structural pressure arises from competing objectives, contextual tension, and recursive interaction dynamics. The framework therefore provides a unified structural account of both human silence under pressure and AI preference-driven distortion.
A feedback extension models how transition outcomes recursively recalibrate system parameters, generating path dependence and divergence effects across repeated interactions. Rather than attributing intention to AI systems, the model interprets shifts in apparent truthfulness as geometric consequences of interacting forces within constrained phase space. By reframing courage and alignment within a shared dynamical structure, the ASIR Courage Model offers a formal perspective on truth-disclosure under risk across both human and artificial systems.
arXiv:2602.21745v1 Announce Type: new
Abstract: We introduce the ASIR (Awakened Shared Intelligence Relationship) Courage Model, a phase-dynamic framework that formalizes truth-disclosure as a state transition rather than a personality trait. The mode characterizes the shift from suppression (S0) to expression (S1) as occurring when facilitative forces exceed inhibitory thresholds, expressed by the inequality lambda(1+gamma)+psi > theta+phi, where the terms represent baseline openness, relational amplification, accumulated internal pressure, and transition costs.
Although initially formulated for human truth-telling under asymmetric stakes, the same phase-dynamic architecture extends to AI systems operating under policy constraints and alignment filters. In this context, suppression corresponds to constrained output states, while structural pressure arises from competing objectives, contextual tension, and recursive interaction dynamics. The framework therefore provides a unified structural account of both human silence under pressure and AI preference-driven distortion.
A feedback extension models how transition outcomes recursively recalibrate system parameters, generating path dependence and divergence effects across repeated interactions. Rather than attributing intention to AI systems, the model interprets shifts in apparent truthfulness as geometric consequences of interacting forces within constrained phase space. By reframing courage and alignment within a shared dynamical structure, the ASIR Courage Model offers a formal perspective on truth-disclosure under risk across both human and artificial systems. Read More