Numerical reasoning over documents, which demands both contextual understanding and logical inference, is challenging for low-capacity local models deployed on computation-constrained devices. Although such complex reasoning queries could be routed to powerful remote models like GPT-4, exposing local data raises significant data leakage concerns. Existing mitigation methods generate problem descriptions or examples for remote assistance. However, the inherent complexity of numerical reasoning hinders the local model from generating logically equivalent queries and accurately inferring answers with remote guidance. In this paper, we present a model collaboration framework with two key innovations: (1) a context-aware synthesis strategy that shifts the query topics while preserving reasoning patterns; and (2) a tool-based answer reconstruction approach that reuses the remote-generated plug-and-play solution with code snippets. Experimental results demonstrate that our method achieves better reasoning accuracy than solely using local models while providing stronger data protection than fully relying on remote models. Furthermore, our method improves accuracy by 16.2% - 43.6% while reducing data leakage by 2.3% - 44.6% compared to existing data protection approaches.

Generating performant executables from high level languages is critical to software performance across a wide range of domains. Modern compilers perform this task by passing code through a series of well-studied optimizations at progressively lower levels of abstraction, but may miss optimization opportunities that require high-level reasoning about a program's purpose. Recent work has proposed using LLMs to fill this gap. While LLMs can achieve large speedups on some programs, they frequently generate code that is incorrect. In this work, we propose a method to balance the correctness of conventional compiler optimizations with the “creativity” of LLM-based code generation: compiler-LLM cooperation. Our approach integrates existing compiler optimization passes with LLM-based code generation at multiple levels of abstraction, retaining the best features of both types of code optimization. We realize our approach with a multi-agent system that includes (1) LLM-based optimization agents for each level of abstraction, (2) individual compiler constituents as tools, (3) an LLM-based test generation agent that probes the correctness and performance of generated code, and (4) a guiding LLM that orchestrates the other components. The strategy enables LLM-based optimization of input programs at multiple levels of abstraction and introduces a method for distributing computational budget between levels. Our extensive evaluation shows that compiler-LLM cooperation outperforms both existing compiler optimizations and level-specific LLM-based baselines, producing speedups up to 1.25x.

Large Language Model (LLM)-based agents have demonstrated remarkable capabilities in complex reasoning and multi-turn interactions but struggle to continuously improve and adapt when deployed in new environments. One promising approach is implementing skill libraries that allow agents to learn, validate, and apply new skills. However, current skill library approaches rely primarily on LLM prompting, making consistent skill library implementation challenging. To overcome these challenges, we propose a Reinforcement Learning (RL)-based approach to enhance agents’ self-improvement capabilities with a skill library. Specifically, we introduce Skill Augmented GRPO for self-Evolution (SAGE), a novel RL framework that systematically incorporates skills into learning. The framework’s key component, Sequential Rollout, iteratively deploys agents across a chain of similar tasks for each rollout. As agents navigate through the task chain, skills generated from previous tasks accumulate in the library and become available for subsequent tasks. Additionally, the framework enhances skill generation and utilization through a Skill-integrated Reward that complements the original outcome-based rewards. Experimental results on AppWorld demonstrate that SAGE, when applied to supervised-finetuned model with expert experience, achieves 8.9% higher Scenario Goal Completion while requiring 26% fewer interaction steps and generating 59% fewer tokens, substantially outperforming existing approaches in both accuracy and efficiency. Our code is available at https://github.com/amazon-science/SAGE.

Large language model agents rely on effective model context to obtain task-relevant information for decision-making. Many existing context engineering approaches primarily rely on the context generated from the past experience and retrieval mechanisms that reuse these context. However, retrieved context from past tasks must be adapted by the execution agent to fit new situations, placing additional reasoning burden on the underlying LLM. To address this limitation, we propose a generative context augmentation framework using Contrastive Learning of Experience via Agentic Reflection (CLEAR). CLEAR first employs a reflection agent to perform contrastive analysis over past execution trajectories and summarize useful context for each observed task. These summaries are then used as supervised fine-tuning data to train a context augmentation model (CAM). Then we further optimize CAM using reinforcement learning, where the reward signal is obtained by running the task execution agent. By learning to generate task-specific knowledge rather than retrieve knowledge from the past, CAM produces context that is better tailored to the current task. We conduct comprehensive evaluations on the AppWorld and WebShop benchmarks. Experimental results show that CLEAR consistently outperforms strong baselines. It improves task completion rate from 72.62

Large Language Models (LLMs) have demonstrated remarkable capabilities, enabling language agents to excel at single-turn tasks. However, their application to complex, multi-step, and long-horizon tasks remains challenging. While reinforcement learning (RL) offers a promising avenue for addressing these challenges, mainstream approaches typically rely solely on sparse, outcome-based rewards, a limitation that becomes especially problematic for group-based RL algorithms lacking critic models, such as Group Relative Policy Optimization (GRPO). In such methods, uniformly rewarding or penalizing all actions within a trajectory can lead to training instability and suboptimal policies, because beneficial and detrimental actions are often entangled across multi-step interactions. To address this challenge, we propose SALT, a novel and lightweight framework that provides a finer-grained advantage assignment, derived solely from outcome rewards. We achieve this by constructing a graph from trajectories of the same prompt, which allows us to quantify the quality of each step and assign advantages accordingly. Crucially, SALT is designed as a plug-and-play module that seamlessly integrates with existing group-based RL algorithms, requiring no modifications to the rollout procedure and introducing negligible computational overhead. Extensive experiments on the WebShop, ALFWorld, and AppWorld benchmarks with various model sizes demonstrate that SALT consistently improves performance. We also conduct a thorough analysis to validate the design choices behind SALT and offer actionable insights.

Since the advent of large language models (LLMs), prompt engineering has been a crucial step for eliciting desired responses for various Natural Language Processing (NLP) tasks. However, prompt engineering remains an impediment for end users due to rapid advances in models, tasks, and associated best practices. To mitigate this, Automatic Prompt Optimization (APO) techniques have recently emerged that use various automated techniques to help improve the performance of LLMs on various tasks. In this paper, we present a comprehensive survey summarizing the current progress and remaining challenges in this field. We provide a formal definition of APO, a 5-part unifying framework, and then proceed to rigorously categorize all relevant works based on their salient features therein. We hope to spur further research guided by our framework.

Vision-Language Models (VLMs) have demonstrated impressive performance across a versatile set of tasks. A key challenge in accelerating VLMs is storing and accessing the large Key-Value (KV) cache that encodes long visual contexts, such as images or videos. While existing KV cache compression methods are effective for Large Language Models (LLMs), directly migrating them to VLMs yields suboptimal accuracy and speedup. To bridge the gap, we propose VL-Cache, a novel KV cache compression recipe tailored for accelerating VLM inference. In this paper, we first investigate the unique sparsity pattern of VLM attention by distinguishing visual and text tokens in prefill and decoding phases. Based on these observations, we introduce a layer-adaptive sparsity-aware cache budget allocation method that effectively distributes the limited cache budget across different layers, further reducing KV cache size without compromising accuracy. Additionally, we develop a modality-aware token scoring policy to better evaluate the token importance. Empirical results on multiple benchmark datasets demonstrate that retaining only 10% of KV cache achieves accuracy comparable to that with full cache. In a speed benchmark, our method accelerates end-to-end latency of generating 100 tokens by up to 2.33x and speeds up decoding by up to 7.08x, while reducing the memory footprint of KV cache in GPU by 90%.

The demand for efficient Large Language Model (LLM) inference has surged with the rising adoption of Generative AI (GenAI) applications, particularly in areas such as agents and retrieval-augmented generation. Efficient inference serves two crucial purposes: it enables the deployment of LLM-centered applications that address critical business needs, while also facilitating rapid experimentation for researchers to extract valuable insights and new understandings. However, despite the field's rapid advancement and interdisciplinary nature, there remains a limited exchange of ideas and methodologies between production-facing practitioners and researchers seeking to experiment with new GenAI concepts quickly. To bridge this gap, we are introducing the first KDD workshop on Inference Optimization for Generative AI. Our goal is to create a collaborative platform where researchers and practitioners working across various use cases and stacks of efficient inference can come together to exchange research ideas, establish connections between different disciplines, and identify challenges and research questions that will shape future work.

Reasoning is increasingly crucial for various tasks. While chain-of-thought prompting enables large language models to leverage reasoning effectively, harnessing the reasoning capabilities of Vision-Language Models (VLMs) remains challenging. To solve this problem, we propose a novel self-distillation framework that enhances the reasoning capabilities of the model. The proposed framework introduces several key innovations. We start by employing a prompt library tailored to visual reasoning tasks to generate diverse in-context questions and utilize a two-step reasoning procedure to derive reasoning-guided responses. These responses are then used for self-distillation, enabling the model to internalize the reasoning process. Additionally, we improve the model architecture with several innovative components, including an intervention adapter for efficient parameter updates, a cross-modal skip connection to facilitate information exchange between modalities, and an ensemble learning algorithm to integrate diverse reasoning from multiple in-context questions. Extensive experiments show that our method significantly improves the baseline performance across five VQA datasets.

Large language models (LLMs) have shown remarkable generalization capability with exceptional performance in various language modeling tasks. However, they still exhibit inherent limitations in precisely capturing and returning grounded knowledge. While existing work has explored utilizing knowledge graphs (KGs) to enhance language modeling via joint training and customized model architectures, applying this to LLMs is problematic owing to their large number of parameters and high computational cost. Therefore, how to enhance pre-trained LLMs using grounded knowledge, e.g., retrieval-augmented generation, remains an open question. In this work, we propose Graph Neural Prompting (GNP), a novel plug-and-play method to assist pre-trained LLMs in learning beneficial knowledge from KGs. GNP encompasses various designs, including a standard graph neural network encoder, a cross-modality pooling module, a domain projector, and a self-supervised link prediction objective. Extensive experiments on multiple datasets demonstrate the superiority of GNP on both commonsense and biomedical reasoning tasks across different LLM sizes and settings. Code is available at https://github.com/meettyj/GNP.

Artificial Intelligence (AI) has enabled breakthroughs on sports analytics in recent years by allowing the extraction of valuable insights and new understandings from vast amounts of sports data. However, due to the highly interdisciplinary nature of the topic, cross-pollination of ideas and methodologies between AI and sports science as well as across different sports remain limited. To address the gap, we propose the first KDD workshop on Data Science and AI for Sports. We aim to bring together researchers and practitioners working on a broad array of sports-related AI/ML use cases, to exchange research ideas, draw connection among various disciplines, and identify challenges and new research questions for future work.

Predictive maintenance (PdM) is one of the most important machine learning (ML) use cases in a wide range of businesses that own or manufacture machinery. Many PdM applications utilize event logs/sequences that record machine operating conditions and health data. However, industrial scale PdM use cases bring a variety of challenges which makes it difficult to directly apply state-of-the-art models. In this work, we start with a PdM use case from a company Light & Wonder (L&W) Inc., which requires us to work with about 280 million events from ∼ 500 gaming machines accumulated over 10 months, and a short 24-hour time window could contain from 0 to approximately 29,000 events. Developing performant models for event sequence data at such scale is a non-trivial task. In this work, we propose a generic, generalizable modeling framework which we refer to as ElasticPdM for industrial scale PdM applications using event log data, which uses novel methods including time-adaptive event sequence binning and time-aware event sequence embedding to process long event sequences. Applied on the real-world L&W data, our ElasticPdM is 13-20% more accurate than state-of-the-art baselines on Recall at High Precision metrics used commonly for PdM.

Classification models learn to generalize the associations between data samples and their target classes. However, researchers have increasingly observed that machine learning practice easily leads to systematic errors in AI applications, a phenomenon referred to as AI blindspots. Such blindspots arise when a model is trained with training samples (e.g., cat/dog classification) where important patterns (e.g., black cats) are missing or periphery/undesirable patterns (e.g., dogs with grass background) are misleading towards a certain class. Even more sophisticated techniques cannot guarantee to capture, reason about, and prevent the spurious associations. In this work, we propose ESCAPE, a visual analytic system that promotes a human-in-the-loop workflow for countering systematic errors. By allowing human users to easily inspect spurious associations, the system facilitates users to spontaneously recognize concepts associated misclassifications and evaluate mitigation strategies that can reduce biased associations. We also propose two statistical approaches, relative concept association to better quantify the associations between a concept and instances, and debias method to mitigate spurious associations. We demonstrate the utility of our proposed ESCAPE system and statistical measures through extensive evaluation including quantitative experiments, usage scenarios, expert interviews, and controlled user experiments.

Machine learning applications are rapidly adopted by industry leaders in any field. The growth of investment in AI-driven solutions created new challenges in managing Data Science and ML resources, people and projects as a whole. The discipline of managing applied machine learning teams, requires a healthy mix between agile product development tool-set and a long term research oriented mindset. The abilities of investing in deep research while at the same time connecting the outcomes to significant business results create a large knowledge based on management methods and best practices in the field. The Second KDD Workshop on Applied Machine Learning Management brings together applied research managers from various fields to share methodologies and case-studies on management of ML teams, products, and projects, achieving business impact with advanced AI-methods.

Proteins perform many essential functions in biological systems and can be successfully developed as bio-therapeutics. It is invaluable to be able to predict their properties based on a proposed sequence and structure. In this study, we developed a novel generalizable deep learning framework, LM-GVP, composed of a protein Language Model (LM) and Graph Neural Network (GNN) to leverage information from both 1D amino acid sequences and 3D structures of proteins. Our approach outperformed the state-of-the-art protein LMs on a variety of property prediction tasks including fluorescence, protease stability, and protein functions from Gene Ontology (GO). We also illustrated insights into how a GNN prediction head can inform the fine-tuning of protein LMs to better leverage structural information. We envision that our deep learning framework will be generalizable to many protein property prediction problems to greatly accelerate protein engineering and drug development.

Proteins perform many essential functions in biological systems and can be successfully developed as bio-therapeutics. It is invaluable to be able to predict their properties based on a proposed sequence and structure. In this study, we developed a novel generalizable deep learning framework, LM-GVP, composed of a protein Language Model (LM) and Graph Neural Network (GNN) to leverage information from both 1D amino acid sequences and 3D structures of proteins. Our approach outperformed the state-of-the-art protein LMs on a variety of property prediction tasks including fluorescence, protease stability, and protein functions from Gene Ontology (GO). We also illustrated insights into how a GNN prediction head can guide the protein LM to better leverage structural information. We envision that our deep learning framework will be generalizable to many protein property prediction problems to greatly accelerate protein engineering and drug development.

The interpretation of deep neural networks (DNNs) has become a key topic as more and more people apply them to solve various problems and making critical decisions. Concept-based explanations have recently become a popular approach for post-hoc interpretation of DNNs. However, identifying human-understandable visual concepts that affect model decisions is a challenging task that is not easily addressed with automatic approaches. We present a novel human-in-the-loop approach to generate user-defined concepts for model interpretation and diagnostics. Central to our proposal is the use of active learning, where human knowledge and feedback are combined to train a concept extractor with very little human labeling effort. We integrate this process into an interactive system, ConceptExtract. Through two case studies, we show how our approach helps analyze model behavior and extract human-friendly concepts for different machine learning tasks and datasets and how to use these concepts to understand the predictions, compare model performance and make suggestions for model refinement. Quantitative experiments show that our active learning approach can accurately extract meaningful visual concepts. More importantly, by identifying visual concepts that negatively affect model performance, we develop the corresponding data augmentation strategy that consistently improves model performance.

Ein Deep-Sequence-Modell mit Prototypen kann gesteuert werden. Eine Prototyp-Ubersicht wird angezeigt, wobei die Prototyp-Ubersicht eine Vielzahl von Prototypsequenzen enthalt, die durch ein Modell durch Ruckwartspropagierung gelernt werden, wobei jede Prototypsequenz eine Reihe von Ereignissen enthalt, wobei fur jede Prototypsequenz statistische Informationen in Bezug auf eine Verwendung der Prototypsequenz durch das Modell aufgezeigt werden. Eine Eingabe wird empfangen, die eine oder mehrere Prototypsequenzen einstellt, um eine Feinabstimmung des Modells vorzunehmen. Das Modell wird unter Verwendung der Vielzahl von Prototypsequenzen aktualisiert, die eingestellt werden, um ein aktualisiertes Modell zu erstellen. Das Modell, das aktualisiert wird, wird in der Prototyp-Ubersicht angezeigt.

Graphs are a ubiquitous data structure to model processes and relations in a wide range of domains. Examples include control-flow graphs in programs and semantic scene graphs in images. Identifying subgraph patterns in graphs is an important approach to understand their structural properties. We propose a visual analytics system GraphQ to support human-in-the-loop, example-based, subgraph pattern search in a database containing many individual graphs. To support fast, interactive queries, we use graph neural networks (GNNs) to encode a graph as fixed-length latent vector representation, and perform subgraph matching in the latent space. Due to the complexity of the problem, it is still difficult to obtain accurate one-to-one node correspondences in the matching results that are crucial for visualization and interpretation. We, therefore, propose a novel GNN for node-alignment called NeuroAlign, to facilitate easy validation and interpretation of the query results. GraphQ provides a visual query interface with a query editor and a multi-scale visualization of the results, as well as a user feedback mechanism for refining the results with additional constraints. We demonstrate GraphQ through two example usage scenarios: analyzing reusable subroutines in program workflows and semantic scene graph search in images. Quantitative experiments show that NeuroAlign achieves 19%-29% improvement in node-alignment accuracy compared to baseline GNN and provides up to 100× speedup compared to combinatorial algorithms. Our qualitative study with domain experts confirms the effectiveness for both usage scenarios.

Ein Verfahren zum Erzeugen einer grafischen Darstellung von zusammengefassten Ereignissequenzen beinhaltet das Empfangen einer Vielzahl von Ereignissequenzen, wobei jede Ereignissequenz in der Vielzahl von Ereignissequenzen eine Vielzahl von Ereignissen umfasst, und dem Erzeugen einer Vielzahl von Clustern unter Verwendung eines Optimierungsprozesses der Mindestbeschreibungslange (MDL). Jeder Cluster in der Vielzahl von Clustern enthalt einen Satz von mindestens zwei Ereignissequenzen in der Vielzahl von Ereignissequenzen, die einem Muster in jedem Cluster zugeordnet sind. Das Muster in jedem Cluster enthalt ferner eine Vielzahl von Ereignissen, die in mindestens einer Ereignissequenz in dem Satz von mindestens zwei Ereignissequenzen in dem Cluster enthalten sind. Das Verfahren beinhaltet das Erzeugen einer grafischen Darstellung eines ersten Clusters in der Vielzahl von Clustern, wobei die grafische Darstellung eine grafische Darstellung einer ersten Vielzahl von Ereignissen in dem Muster des ersten Clusters enthalt.