AgentDS Technical Report: Benchmarking the Future of Human-AI Collaboration in Domain-Specific Data Science

Data science plays a critical role in transforming complex data into actionable insights across numerous domains. Recent developments in large language models (LLMs) and artificial intelligence (AI) agents have significantly automated data science workflow. However, it remains unclear to what extent AI agents can match the performance of human experts on domain-specific data science tasks, and in which aspects human expertise continues to provide advantages. We introduce AgentDS, a benchmark and competition designed to evaluate both AI agents and human-AI collaboration performance in domain-specific data science. AgentDS consists of 17 challenges across six industries: commerce, food production, healthcare, insurance, manufacturing, and retail banking. We conducted an open competition involving 29 teams and 80 participants, enabling systematic comparison between human-AI collaborative approaches and AI-only baselines. Our results show that current AI agents struggle with domain-specific reasoning. AI-only baselines perform near or below the median of competition participants, while the strongest solutions arise from human-AI collaboration. These findings challenge the narrative of complete automation by AI and underscore the enduring importance of human expertise in data science, while illuminating directions for the next generation of AI. Visit the AgentDS website here: https://agentds.org/ and open source datasets here: https://huggingface.co/datasets/lainmn/AgentDS .

StitchCUDA: An Automated Multi-Agents End-to-End GPU Programing Framework with Rubric-based Agentic Reinforcement Learning

Modern machine learning (ML) workloads increasingly rely on GPUs, yet achieving high end-to-end performance remains challenging due to dependencies on both GPU kernel efficiency and host-side settings. Although LLM-based methods show promise on automated GPU kernel generation, prior works mainly focus on single-kernel optimization and do not extend to end-to-end programs, hindering practical deployment. To address the challenge, in this work, we propose StitchCUDA, a multi-agent framework for end-to-end GPU program generation, with three specialized agents: a Planner to orchestrate whole system design, a Coder dedicated to implementing it step-by-step, and a Verifier for correctness check and performance profiling using Nsys/NCU. To fundamentally improve the Coder's ability in end-to-end GPU programming, StitchCUDA integrates rubric-based agentic reinforcement learning over two atomic skills, task-to-code generation and feedback-driven code optimization, with combined rubric reward and rule-based reward from real executions. Therefore, the Coder learns how to implement advanced CUDA programming techniques (e.g., custom kernel fusion, cublas epilogue), and we also effectively prevent Coder's reward hacking (e.g., just copy PyTorch code or hardcoding output) during benchmarking. Experiments on KernelBench show that StitchCUDA achieves nearly 100% success rate on end-to-end GPU programming tasks, with 1.72x better speedup over the multi-agent baseline and 2.73x than the RL model baselines.

Powering Up Zeroth-Order Training via Subspace Gradient Orthogonalization

Zeroth-order (ZO) optimization provides a gradient-free alternative to first-order (FO) methods by estimating gradients via finite differences of function evaluations, and has recently emerged as a memory-efficient paradigm for fine-tuning large-scale models by avoiding backpropagation. However, ZO optimization has a fundamental tension between accuracy and query efficiency. In this work, we show that ZO optimization can be substantially improved by unifying two complementary principles: (i) a projection-based subspace view that reduces gradient estimation variance by exploiting the intrinsic low-rank structure of model updates, and (ii) Muon-style spectral optimization that applies gradient orthogonalization to extract informative spectral structure from noisy ZO gradients. These findings form a unified framework of subspace gradient orthogonalization, which we instantiate in a new method, ZO-Muon, admitting a natural interpretation as a low-rank Muon optimizer in the ZO setting. Extensive experiments on large language models (LLMs) and vision transformers (ViTs) demonstrate that ZO-Muon significantly accelerates convergence and achieves a win-win improvement in accuracy and query/runtime efficiency. Notably, compared to the popular MeZO baseline, ZO-Muon requires only 24.7% of the queries to reach the same SST-2 performance for LLM fine-tuning, and improves accuracy by 25.1% on ViT-B fine-tuning on CIFAR-100.

HiPER: Hierarchical Reinforcement Learning with Explicit Credit Assignment for Large Language Model Agents

Training LLMs as interactive agents for multi-turn decision-making remains challenging, particularly in long-horizon tasks with sparse and delayed rewards, where agents must execute extended sequences of actions before receiving meaningful feedback. Most existing reinforcement learning (RL) approaches model LLM agents as flat policies operating at a single time scale, selecting one action at each turn. In sparse-reward settings, such flat policies must propagate credit across the entire trajectory without explicit temporal abstraction, which often leads to unstable optimization and inefficient credit assignment. We propose HiPER, a novel Hierarchical Plan-Execute RL framework that explicitly separates high-level planning from low-level execution. HiPER factorizes the policy into a high-level planner that proposes subgoals and a low-level executor that carries them out over multiple action steps. To align optimization with this structure, we introduce a key technique called hierarchical advantage estimation (HAE), which carefully assigns credit at both the planning and execution levels. By aggregating returns over the execution of each subgoal and coordinating updates across the two levels, HAE provides an unbiased gradient estimator and provably reduces variance compared to flat generalized advantage estimation. Empirically, HiPER achieves state-of-the-art performance on challenging interactive benchmarks, reaching 97.4\% success on ALFWorld and 83.3\% on WebShop with Qwen2.5-7B-Instruct (+6.6\% and +8.3\% over the best prior method), with especially large gains on long-horizon tasks requiring multiple dependent subtasks. These results highlight the importance of explicit hierarchical decomposition for scalable RL training of multi-turn LLM agents.

DISPO: Enhancing Training Efficiency and Stability in Reinforcement Learning for Large Language Model Mathematical Reasoning

Reinforcement learning with verifiable rewards has emerged as a promising paradigm for enhancing the reasoning capabilities of large language models particularly in mathematics. Current approaches in this domain present a clear trade-off: PPO-style methods (e.g., GRPO/DAPO) offer training stability but exhibit slow learning trajectories due to their trust-region constraints on policy updates, while REINFORCE-style approaches (e.g., CISPO) demonstrate improved learning efficiency but suffer from performance instability as they clip importance sampling weights while still permitting non-zero gradients outside the trust-region. To address these limitations, we introduce DISPO, a simple yet effective REINFORCE-style algorithm that decouples the up-clipping and down-clipping of importance sampling weights for correct and incorrect responses, yielding four controllable policy update regimes. Through targeted ablations, we uncover how each regime impacts training: for correct responses, weights >1 increase the average token entropy (i.e., exploration) while weights <1 decrease it (i.e., distillation) -- both beneficial but causing gradual performance degradation when excessive. For incorrect responses, overly restrictive clipping triggers sudden performance collapse through repetitive outputs (when weights >1) or vanishing response lengths (when weights <1). By separately tuning these four clipping parameters, DISPO maintains the exploration-distillation balance while preventing catastrophic failures, achieving 61.04% on AIME'24 (vs. 55.42% CISPO and 50.21% DAPO) with similar gains across various benchmarks and models.

Scaling Unverifiable Rewards: A Case Study on Visual Insights

Large Language Model (LLM) agents can increasingly automate complex reasoning through Test-Time Scaling (TTS), iterative refinement guided by reward signals. However, many real-world tasks involve multi-stage pipeline whose final outcomes lack verifiable rewards or sufficient data to train robust reward models, making judge-based refinement prone to accumulate error over stages. We propose Selective TTS, a process-based refinement framework that scales inference across different stages in multi-agent pipeline, instead of repeated refinement over time by prior work. By distributing compute across stages and pruning low-quality branches early using process-specific judges, Selective TTS mitigates the judge drift and stabilizes refinement. Grounded in the data science pipeline, we build an end-to-end multi-agent pipeline for generating visually insightful charts and report of given dataset, and design a reliable LLM-based judge model, aligned with human experts (Kendall's τ=0.55). Our proposed selective TTS then improves insight quality under a fixed compute budget, increasing mean scores from 61.64 to 65.86 while reducing variance. We hope our findings serve as the first step toward to scaling complex, open-ended tasks with unverifiable rewards, such as scientific discovery and story generation.

Can Agentic AI Match the Performance of Human Data Scientists?

Data science plays a critical role in transforming complex data into actionable insights across numerous domains. Recent developments in large language models (LLMs) have significantly automated data science workflows, but a fundamental question persists: Can these agentic AI systems truly match the performance of human data scientists who routinely leverage domain-specific knowledge? We explore this question by designing a prediction task where a crucial latent variable is hidden in relevant image data instead of tabular features. As a result, agentic AI that generates generic codes for modeling tabular data cannot perform well, while human experts could identify the important hidden variable using domain knowledge. We demonstrate this idea with a synthetic dataset for property insurance. Our experiments show that agentic AI that relies on generic analytics workflow falls short of methods that use domain-specific insights. This highlights a key limitation of the current agentic AI for data science and underscores the need for future research to develop agentic AI systems that can better recognize and incorporate domain knowledge.

Leak@$k$: Unlearning Does Not Make LLMs Forget Under Probabilistic Decoding

Unlearning in large language models (LLMs) is critical for regulatory compliance and for building ethical generative AI systems that avoid producing private, toxic, illegal, or copyrighted content. Despite rapid progress, in this work we show that \textit{almost all} existing unlearning methods fail to achieve true forgetting in practice. Specifically, while evaluations of these `unlearned' models under deterministic (greedy) decoding often suggest successful knowledge removal using standard benchmarks (as has been done in the literature), we show that sensitive information reliably resurfaces when models are sampled with standard probabilistic decoding. To rigorously capture this vulnerability, we introduce \texttt{leak@$k$}, a new meta-evaluation metric that quantifies the likelihood of forgotten knowledge reappearing when generating $k$ samples from the model under realistic decoding strategies. Using three widely adopted benchmarks, TOFU, MUSE, and WMDP, we conduct the first large-scale, systematic study of unlearning reliability using our newly defined \texttt{leak@$k$} metric. Our findings demonstrate that knowledge leakage persists across methods and tasks, underscoring that current state-of-the-art unlearning techniques provide only limited forgetting and highlighting the urgent need for more robust approaches to LLM unlearning.

A Framework for Quantifying How Pre-Training and Context Benefit In-Context Learning

Pre-trained large language models have demonstrated a strong ability to learn from context, known as in-context learning (ICL). Despite a surge of recent applications that leverage such capabilities, it is by no means clear, at least theoretically, how the ICL capabilities arise, and in particular, what is the precise role played by key factors such as pre-training procedure as well as context construction. In this work, we propose a new framework to analyze the ICL performance, for a class of realistic settings, which includes network architectures, data encoding, data generation, and prompt construction process. As a first step, we construct a simple example with a one-layer transformer, and show an interesting result, namely when the pre-train data distribution is different from the query task distribution, a properly constructed context can shift the output distribution towards the query task distribution, in a quantifiable manner, leading to accurate prediction on the query topic. We then extend the findings in the previous step to a more general case, and derive the precise relationship between ICL performance, context length and the KL divergence between pre-train and query task distribution. Finally, we provide experiments to validate our theoretical results.

BLUR: A Bi-Level Optimization Approach for LLM Unlearning

Enabling large language models (LLMs) to unlearn knowledge and capabilities acquired during training has proven vital for ensuring compliance with data regulations and promoting ethical practices in generative AI. Although there are growing interests in developing various unlearning algorithms, it remains unclear how to best formulate the unlearning problem. The most popular formulation uses a weighted sum of forget and retain loss, but it often leads to performance degradation due to the inherent trade-off between forget and retain losses. In this work, we argue that it is important to model the hierarchical structure of the unlearning problem, where the forget problem (which \textit{unlearns} certain knowledge and/or capabilities) takes priority over the retain problem (which preserves model utility). This hierarchical structure naturally leads to a bi-level optimization formulation where the lower-level objective focuses on minimizing the forget loss, while the upper-level objective aims to maintain the model's utility. Based on this new formulation, we propose a novel algorithm, termed Bi-Level UnleaRning (\texttt{BLUR}), which not only possesses strong theoretical guarantees but more importantly, delivers superior performance. In particular, our extensive experiments demonstrate that \texttt{BLUR} consistently outperforms all the state-of-the-art algorithms across various unlearning tasks, models, and metrics. Codes are available at https://github.com/OptimAI-Lab/BLURLLMUnlearning.