Large Language Models (LLMs) offer a promising basis for creating agents that can tackle complex tasks through iterative environmental interaction. Existing methods either require these agents to mimic expert-provided trajectories or rely on definitive environmental feedback for reinforcement learning which limits their application to specific scenarios like gaming or code generation. This paper introduces a novel training method for LLM-based agents using weakly supervised signals from a critic LLM, bypassing the need for expert trajectories or definitive feedback. Our agents are trained in iterative manner, where they initially generate trajectories through environmental interaction. Subsequently, a critic LLM selects a subset of good trajectories, which are then used to update the agents, enabling them to generate improved trajectories in the next iteration. Extensive tests on the API-bank dataset show consistent improvement in our agents' capabilities and comparable performance to GPT-4, despite using open-source models with much fewer parameters.
Flow-matching models provide a powerful framework for various applications, offering efficient sampling and flexible probability path modeling. These models are characterized by flows with low curvature in learned generative trajectories, which results in reduced truncation error at each sampling step. To further reduce curvature, we propose block matching. This novel approach leverages label information to partition the data distribution into blocks and match them with a prior distribution parameterized using the same label information, thereby learning straighter flows. We demonstrate that the variance of the prior distribution can control the curvature upper bound of forward trajectories in flow-matching models. By designing flexible regularization strategies to adjust this variance, we achieve optimal generation performance, effectively balancing the trade-off between maintaining diversity in generated samples and minimizing numerical solver errors. Our results demonstrate competitive performance with models of the same parameter scale.Code is available at \url{https://github.com/wpp13749/block_flow}.
We introduce QualityFlow, a dynamic agentic workflow for program synthesis. Given the English description of a programming problem and a set of unit tests, the model's goal is to synthesize the correct program that solves the problem and passes the tests. QualityFlow includes large language model (LLM) agents resembling a software development team, including code generation, testing, and self-debugging. We propose the LLM Quality Checker, which explicitly "imagines" whether the synthesized programs' execution would conform to the unit tests. The Quality Checks dynamically control the workflow, including actions to submit the final answer, clarify the problem statement, and revert previous workflow steps. Our experiments show that the Quality Checker can precisely accept any correct program, mitigate faulty synthesized tests, and prevent potential workflow deviation. QualityFlow establishes the state-of-the-art results on four program synthesis benchmarks: MBPP, HumanEval, and stricter evaluations from MBPP-EvalPlus and HumanEval-EvalPlus.
In recent years, numerous neural network architectures specifically designed for the instance segmentation of nuclei in microscopic images have been released. These models embed nuclei-specific priors to outperform generic architectures like U-Nets; however, they require large annotated datasets, which are often not available. Generative models (GANs, diffusion models) have been used to compensate for this by synthesizing training data. These two-stage approaches are computationally expensive, as first a generative model and then a segmentation model has to be trained. We propose CyclePose, a hybrid framework integrating synthetic data generation and segmentation training. CyclePose builds on a CycleGAN architecture, which allows unpaired translation between microscopy images and segmentation masks. We embed a segmentation model into CycleGAN and leverage a cycle consistency loss for self-supervision. Without annotated data, CyclePose outperforms other weakly or unsupervised methods on two public datasets. Code is available at https://github.com/jonasutz/CyclePose
To address the challenge of insufficient interactivity and behavioral diversity in autonomous driving decision-making, this paper proposes a Cognitive Hierarchical Agent for Reasoning and Motion Stylization (CHARMS). By leveraging Level-k game theory, CHARMS captures human-like reasoning patterns through a two-stage training pipeline comprising reinforcement learning pretraining and supervised fine-tuning. This enables the resulting models to exhibit diverse and human-like behaviors, enhancing their decision-making capacity and interaction fidelity in complex traffic environments. Building upon this capability, we further develop a scenario generation framework that utilizes the Poisson cognitive hierarchy theory to control the distribution of vehicles with different driving styles through Poisson and binomial sampling. Experimental results demonstrate that CHARMS is capable of both making intelligent driving decisions as an ego vehicle and generating diverse, realistic driving scenarios as environment vehicles. The code for CHARMS is released at https://github.com/chuduanfeng/CHARMS.
Large Language Models (LLMs) represent a landmark achievement in Artificial Intelligence (AI), demonstrating unprecedented proficiency in procedural tasks such as text generation, code completion, and conversational coherence. These capabilities stem from their architecture, which mirrors human procedural memory -- the brain's ability to automate repetitive, pattern-driven tasks through practice. However, as LLMs are increasingly deployed in real-world applications, it becomes impossible to ignore their limitations operating in complex, unpredictable environments. This paper argues that LLMs, while transformative, are fundamentally constrained by their reliance on procedural memory. To create agents capable of navigating ``wicked'' learning environments -- where rules shift, feedback is ambiguous, and novelty is the norm -- we must augment LLMs with semantic memory and associative learning systems. By adopting a modular architecture that decouples these cognitive functions, we can bridge the gap between narrow procedural expertise and the adaptive intelligence required for real-world problem-solving.
Large Language Models (LLMs) have become foundational in modern artificial intelligence, powering a wide range of applications from code generation and virtual assistants to scientific research and enterprise automation. However, concerns about data contamination--where test data overlaps with training data--have raised serious questions about the reliability of these applications. Despite awareness of this issue, existing methods fall short in effectively identifying or mitigating contamination. In this paper, we propose Residual-Noise Fingerprinting (RN-F), a novel framework for detecting contaminated data in LLMs. RN-F is a single-pass, gradient-free detection method that leverages residual signal patterns without introducing additional floating-point operations. Our approach is lightweight, model-agnostic, and efficient. We evaluate RN-F on multiple LLMs across various contaminated datasets and show that it consistently outperforms existing state-of-the-art methods, achieving performance improvements of up to 10.5% in contamination detection metrics.
In recent years, multi-agent frameworks powered by large language models (LLMs) have advanced rapidly. Despite this progress, there is still a notable absence of benchmark datasets specifically tailored to evaluate their performance. To bridge this gap, we introduce Auto-SLURP, a benchmark dataset aimed at evaluating LLM-based multi-agent frameworks in the context of intelligent personal assistants. Auto-SLURP extends the original SLURP dataset -- initially developed for natural language understanding tasks -- by relabeling the data and integrating simulated servers and external services. This enhancement enables a comprehensive end-to-end evaluation pipeline, covering language understanding, task execution, and response generation. Our experiments demonstrate that Auto-SLURP presents a significant challenge for current state-of-the-art frameworks, highlighting that truly reliable and intelligent multi-agent personal assistants remain a work in progress. The dataset and related code are available at https://github.com/lorashen/Auto-SLURP/.
Large language models (LLMs) have demonstrated exceptional performance in planning the use of various functional tools, such as calculators and retrievers, particularly in question-answering tasks. In this paper, we expand the definition of these tools, centering on conceptual tools within the context of dialogue systems. A conceptual tool specifies a cognitive concept that aids systematic or investigative thought. These conceptual tools play important roles in practice, such as multiple psychological or tutoring strategies being dynamically applied in a single turn to compose helpful responses. To further enhance the reasoning and planning capability of LLMs with these conceptual tools, we introduce a multi-persona collaboration framework: Think-Plan-Execute (TPE). This framework decouples the response generation process into three distinct roles: Thinker, Planner, and Executor. Specifically, the Thinker analyzes the internal status exhibited in the dialogue context, such as user emotions and preferences, to formulate a global guideline. The Planner then generates executable plans to call different conceptual tools (e.g., sources or strategies), while the Executor compiles all intermediate results into a coherent response. This structured approach not only enhances the explainability and controllability of responses but also reduces token redundancy. We demonstrate the effectiveness of TPE across various dialogue response generation tasks, including multi-source (FoCus) and multi-strategy interactions (CIMA and PsyQA). This reveals its potential to handle real-world dialogue interactions that require more complicated tool learning beyond just functional tools. The full code and data will be released for reproduction.
As Large Language Models (LLMs) have made significant advancements across various tasks, such as question answering, translation, text summarization, and dialogue systems, the need for accuracy in information becomes crucial, especially for serious financial products serving billions of users like Alipay. However, for a real-world product serving millions of users, the inference speed of LLMs becomes a critical factor compared to a mere experimental model. Hence, this paper presents a generic framework for accelerating the inference process, resulting in a substantial increase in speed and cost reduction for our LLM-based scenarios, with lossless generation accuracy. In the traditional inference process, each token is generated sequentially by the LLM, leading to a time consumption proportional to the number of generated tokens. To enhance this process, our framework, named \textit{lookahead}, introduces a \textit{multi-branch} strategy. Instead of generating a single token at a time, we propose a Trie-based retrieval and verification mechanism to be able to accept several tokens at a forward step. Our strategy offers two distinct advantages: (1) it guarantees absolute correctness of the output, avoiding any approximation algorithms, and (2) the worst-case performance of our approach is equivalent to the conventional process. We conduct extensive experiments to demonstrate the significant improvements achieved by applying our inference acceleration framework. Our framework is widely deployed in Alipay since April 2023, and obtain remarkable 2.66x to 6.26x speedup. Our code is available at https://github.com/alipay/PainlessInferenceAcceleration.
Masked diffusion models (MDMs) have shown promise in language modeling, yet their scalability and effectiveness in core language tasks, such as text generation and language understanding, remain underexplored. This paper establishes the first scaling law for MDMs, demonstrating a scaling rate comparable to autoregressive models (ARMs) and a relatively small compute gap. Motivated by their scalability, we train a family of MDMs with up to 1.1 billion (B) parameters to systematically evaluate their performance against ARMs of comparable or larger sizes. Fully leveraging the probabilistic formulation of MDMs, we propose a simple yet effective unsupervised classifier-free guidance that effectively exploits large-scale unpaired data, boosting performance for conditional inference. In language understanding, the 1.1B MDM outperforms the 1.1B TinyLlama model trained on the same data across four of eight zero-shot benchmarks. Notably, it achieves competitive math reasoning ability with the 7B Llama-2 model on the GSM8K dataset. In text generation, MDMs provide a flexible trade-off compared to ARMs utilizing KV-cache: MDMs match the performance of ARMs while being 1.4 times faster or achieving higher quality than ARMs at a higher computational cost. Moreover, MDMs address challenging tasks for ARMs by effectively handling bidirectional reasoning and adapting to temporal shifts in data. Notably, a 1.1B MDM breaks the reverse curse encountered by much larger ARMs with significantly more data and computation, such as 13B Llama-2 and 175B GPT-3. Our code is available at https://github.com/ML-GSAI/SMDM.
Software development is shifting from traditional logical programming to model-integrated applications that leverage generative AI and large language models (LLMs) during runtime. However, integrating LLMs remains complex, requiring developers to manually craft prompts and process outputs. Existing tools attempt to assist with prompt engineering, but often introduce additional complexity. This paper presents Meaning-Typed Programming (MTP) model, a novel paradigm that abstracts LLM integration through intuitive language-level constructs. By leveraging the inherent semantic richness of code, MTP automates prompt generation and response handling without additional developer effort. We introduce the by operator for seamless LLM invocation, MT-IR, a meaning-based intermediate representation for semantic extraction, and MT-Runtime, an automated system for managing LLM interactions. We implement MTP in Jac, a Python superset language and find that MTP significantly reduces coding complexity while maintaining accuracy and efficiency. Our evaluation across diverse benchmarks and user studies demonstrates that MTP outperforms existing frameworks such as DSPy and LMQL by reducing lines of code by factors of 2.3-7.5X and 1.3-10.7X respectively. For math problems from the GSM8k dataset, MTP achieves accuracy rates approaching 90%, while reducing token usage in 10 out of 13 benchmarks. This leads to cost savings up to 4.5X and runtime speedups as high as 4.75X. Additionally, MTP demonstrates resilience even when 50% of naming conventions are suboptimal, establishing it as a practical, efficient solution for streamlining model-integrated application development.
Recent advancements in software engineering agents have demonstrated promising capabilities in automating program improvements. However, their reliance on closed-source or resource-intensive models introduces significant deployment challenges in private environments, prompting a critical question: \textit{How can personally deployable open-source LLMs achieve comparable code reasoning performance?} To this end, we propose a unified Test-Time Compute scaling framework that leverages increased inference-time computation instead of larger models. Our framework incorporates two complementary strategies: internal TTC and external TTC. Internally, we introduce a \textit{development-contextualized trajectory synthesis} method leveraging real-world software repositories to bootstrap multi-stage reasoning processes, such as fault localization and patch generation. We further enhance trajectory quality through rejection sampling, rigorously evaluating trajectories along accuracy and complexity. Externally, we propose a novel \textit{development-process-based search} strategy guided by reward models and execution verification. This approach enables targeted computational allocation at critical development decision points, overcoming limitations of existing "end-point only" verification methods. Evaluations on SWE-bench Verified demonstrate our \textbf{32B model achieves a 46\% issue resolution rate}, surpassing significantly larger models such as DeepSeek R1 671B and OpenAI o1. Additionally, we provide the empirical validation of the test-time scaling phenomenon within SWE agents, revealing that \textbf{models dynamically allocate more tokens to increasingly challenging problems}, effectively enhancing reasoning capabilities. We publicly release all training data, models, and code to facilitate future research. https://github.com/yingweima2022/SWE-Reasoner
We introduce AgentAda, the first LLM-powered analytics agent that can learn and use new analytics skills to extract more specialized insights. Unlike existing methods that require users to manually decide which data analytics method to apply, AgentAda automatically identifies the skill needed from a library of analytical skills to perform the analysis. This also allows AgentAda to use skills that existing LLMs cannot perform out of the box. The library covers a range of methods, including clustering, predictive modeling, and NLP techniques like BERT, which allow AgentAda to handle complex analytics tasks based on what the user needs. AgentAda's dataset-to-insight extraction strategy consists of three key steps: (I) a question generator to generate queries relevant to the user's goal and persona, (II) a hybrid Retrieval-Augmented Generation (RAG)-based skill matcher to choose the best data analytics skill from the skill library, and (III) a code generator that produces executable code based on the retrieved skill's documentation to extract key patterns. We also introduce KaggleBench, a benchmark of curated notebooks across diverse domains, to evaluate AgentAda's performance. We conducted a human evaluation demonstrating that AgentAda provides more insightful analytics than existing tools, with 48.78% of evaluators preferring its analyses, compared to 27.67% for the unskilled agent. We also propose a novel LLM-as-a-judge approach that we show is aligned with human evaluation as a way to automate insight quality evaluation at larger scale.
We present AVOD, an Aggregate View Object Detection network for autonomous driving scenarios. The proposed neural network architecture uses LIDAR point clouds and RGB images to generate features that are shared by two subnetworks: a region proposal network (RPN) and a second stage detector network. The proposed RPN uses a novel architecture capable of performing multimodal feature fusion on high resolution feature maps to generate reliable 3D object proposals for multiple object classes in road scenes. Using these proposals, the second stage detection network performs accurate oriented 3D bounding box regression and category classification to predict the extents, orientation, and classification of objects in 3D space. Our proposed architecture is shown to produce state of the art results on the KITTI 3D object detection benchmark while running in real time with a low memory footprint, making it a suitable candidate for deployment on autonomous vehicles. Code is at: https://github.com/kujason/avod
Generating and manipulating human facial images using high-level attributal controls are important and interesting problems. The models proposed in previous work can solve one of these two problems (generation or manipulation), but not both coherently. This paper proposes a novel model that learns how to both generate and modify the facial image from high-level semantic attributes. Our key idea is to formulate a Semi-Latent Facial Attribute Space (SL-FAS) to systematically learn relationship between user-defined and latent attributes, as well as between those attributes and RGB imagery. As part of this newly formulated space, we propose a new model --- SL-GAN which is a specific form of Generative Adversarial Network. Finally, we present an iterative training algorithm for SL-GAN. The experiments on recent CelebA and CASIA-WebFace datasets validate the effectiveness of our proposed framework. We will also make data, pre-trained models and code available.
Traditional image recognition methods only consider objects belonging to already learned classes. However, since training a recognition model with every object class in the world is unfeasible, a way of getting information on unknown objects (i.e., objects whose class has not been learned) is necessary. A way for an image recognition system to learn new classes could be asking a human about objects that are unknown. In this paper, we propose a method for generating questions about unknown objects in an image, as means to get information about classes that have not been learned. Our method consists of a module for proposing objects, a module for identifying unknown objects, and a module for generating questions about unknown objects. The experimental results via human evaluation show that our method can successfully get information about unknown objects in an image dataset. Our code and dataset are available at https://github.com/mil-tokyo/vqg-unknown.
People enjoy food photography because they appreciate food. Behind each meal there is a story described in a complex recipe and, unfortunately, by simply looking at a food image we do not have access to its preparation process. Therefore, in this paper we introduce an inverse cooking system that recreates cooking recipes given food images. Our system predicts ingredients as sets by means of a novel architecture, modeling their dependencies without imposing any order, and then generates cooking instructions by attending to both image and its inferred ingredients simultaneously. We extensively evaluate the whole system on the large-scale Recipe1M dataset and show that (1) we improve performance w.r.t. previous baselines for ingredient prediction; (2) we are able to obtain high quality recipes by leveraging both image and ingredients; (3) our system is able to produce more compelling recipes than retrieval-based approaches according to human judgment. We make code and models publicly available.
We explore different design choices for injecting noise into generative adversarial networks (GANs) with the goal of disentangling the latent space. Instead of traditional approaches, we propose feeding multiple noise codes through separate fully-connected layers respectively. The aim is restricting the influence of each noise code to specific parts of the generated image. We show that disentanglement in the first layer of the generator network leads to disentanglement in the generated image. Through a grid-based structure, we achieve several aspects of disentanglement without complicating the network architecture and without requiring labels. We achieve spatial disentanglement, scale-space disentanglement, and disentanglement of the foreground object from the background style allowing fine-grained control over the generated images. Examples include changing facial expressions in face images, changing beak length in bird images, and changing car dimensions in car images. This empirically leads to better disentanglement scores than state-of-the-art methods on the FFHQ dataset.
Joint entity and relation extraction aims to extract relation triplets from plain text directly. Prior work leverages Sequence-to-Sequence (Seq2Seq) models for triplet sequence generation. However, Seq2Seq enforces an unnecessary order on the unordered triplets and involves a large decoding length associated with error accumulation. These introduce exposure bias, which may cause the models overfit to the frequent label combination, thus deteriorating the generalization. We propose a novel Sequence-to-Unordered-Multi-Tree (Seq2UMTree) model to minimize the effects of exposure bias by limiting the decoding length to three within a triplet and removing the order among triplets. We evaluate our model on two datasets, DuIE and NYT, and systematically study how exposure bias alters the performance of Seq2Seq models. Experiments show that the state-of-the-art Seq2Seq model overfits to both datasets while Seq2UMTree shows significantly better generalization. Our code is available at https://github.com/WindChimeRan/OpenJERE .