arXiv subjects

Yunhao Yang

Publications and source records attributed to Yunhao Yang.

34 records · Page 2Linked to original sources

arXiv2024

Reasoning, Memorization, and Fine-Tuning Language Models for Non-Cooperative Games

We develop a method that integrates the tree of thoughts and multi-agent framework to enhance the capability of pre-trained language models in solving complex, unfamiliar games. The method decomposes game-solving into four incremental tasks -- game summarization, area selection, action extraction, and action validation -- each assigned to a specific language-model agent. By constructing a tree of thoughts, the method simulates reasoning paths and allows agents to collaboratively distill game representations and tactics, mitigating the limitations of language models in reasoning and long-term memorization. Additionally, an automated fine-tuning process further optimizes the agents' performance by ranking query-response pairs based on game outcomes, e.g., winning or losing. We apply the method to a non-cooperative game and demonstrate a 65 percent winning rate against benchmark algorithms, with an additional 10 percent improvement after fine-tuning. In contrast to existing deep learning algorithms for game solving that require millions of training samples, the proposed method consumes approximately 1000 training samples, highlighting its efficiency and scalability.

cs.AI↗

arXiv2024

Uncertainty-Guided Enhancement on Driving Perception System via Foundation Models

Multimodal foundation models offer promising advancements for enhancing driving perception systems, but their high computational and financial costs pose challenges. We develop a method that leverages foundation models to refine predictions from existing driving perception models -- such as enhancing object classification accuracy -- while minimizing the frequency of using these resource-intensive models. The method quantitatively characterizes uncertainties in the perception model's predictions and engages the foundation model only when these uncertainties exceed a pre-specified threshold. Specifically, it characterizes uncertainty by calibrating the perception model's confidence scores into theoretical lower bounds on the probability of correct predictions using conformal prediction. Then, it sends images to the foundation model and queries for refining the predictions only if the theoretical bound of the perception model's outcome is below the threshold. Additionally, we propose a temporal inference mechanism that enhances prediction accuracy by integrating historical predictions, leading to tighter theoretical bounds. The method demonstrates a 10 to 15 percent improvement in prediction accuracy and reduces the number of queries to the foundation model by 50 percent, based on quantitative evaluations from driving datasets.

cs.CV↗

arXiv2024

Multimodal Pretrained Models for Verifiable Sequential Decision-Making: Planning, Grounding, and Perception

Recently developed pretrained models can encode rich world knowledge expressed in multiple modalities, such as text and images. However, the outputs of these models cannot be integrated into algorithms to solve sequential decision-making tasks. We develop an algorithm that utilizes the knowledge from pretrained models to construct and verify controllers for sequential decision-making tasks, and to ground these controllers to task environments through visual observations with formal guarantees. In particular, the algorithm queries a pretrained model with a user-provided, text-based task description and uses the model's output to construct an automaton-based controller that encodes the model's task-relevant knowledge. It allows formal verification of whether the knowledge encoded in the controller is consistent with other independently available knowledge, which may include abstract information on the environment or user-provided specifications. Next, the algorithm leverages the vision and language capabilities of pretrained models to link the observations from the task environment to the text-based control logic from the controller (e.g., actions and conditions that trigger the actions). We propose a mechanism to provide probabilistic guarantees on whether the controller satisfies the user-provided specifications under perceptual uncertainties. We demonstrate the algorithm's ability to construct, verify, and ground automaton-based controllers through a suite of real-world tasks, including daily life and robot manipulation tasks.

cs.AI↗

arXiv2023

Fine-Tuning Language Models Using Formal Methods Feedback

Although pre-trained language models encode generic knowledge beneficial for planning and control, they may fail to generate appropriate control policies for domain-specific tasks. Existing fine-tuning methods use human feedback to address this limitation, however, sourcing human feedback is labor intensive and costly. We present a fully automated approach to fine-tune pre-trained language models for applications in autonomous systems, bridging the gap between generic knowledge and domain-specific requirements while reducing cost. The method synthesizes automaton-based controllers from pre-trained models guided by natural language task descriptions. These controllers are verifiable against independently provided specifications within a world model, which can be abstract or obtained from a high-fidelity simulator. Controllers with high compliance with the desired specifications receive higher ranks, guiding the iterative fine-tuning process. We provide quantitative evidences, primarily in autonomous driving, to demonstrate the method's effectiveness across multiple tasks. The results indicate an improvement in percentage of specifications satisfied by the controller from 60% to 90%.

cs.AI↗

arXiv2023

Deep Contrastive Patch-Based Subspace Learning for Camera Image Signal Processing

Camera Image Signal Processing (ISP) pipelines can get appealing results in different image signal processing tasks. Nonetheless, the majority of these methods, including those employing an encoder-decoder deep architecture for the task, typically utilize a uniform filter applied consistently across the entire image. However, it is natural to view a camera image as heterogeneous, as the color intensity and the artificial noise are distributed vastly differently, even across the two-dimensional domain of a single image. Varied Moire ringing, motion blur, color-bleaching, or lens-based projection distortions can all potentially lead to a heterogeneous image artifact filtering problem. In this paper, we present a specific patch-based, local subspace deep neural network that improves Camera ISP to be robust to heterogeneous artifacts (especially image denoising). We call our three-fold deep-trained model the Patch Subspace Learning Autoencoder (PSL-AE). The PSL-AE model does not make assumptions regarding uniform levels of image distortion. Instead, it first encodes patches extracted from noisy a nd clean image pairs, with different artifact types or distortion levels, by contrastive learning. Then, the patches of each image are encoded into corresponding soft clusters within their suitable latent sub-space, utilizing a prior mixture model. Furthermore, the decoders undergo training in an unsupervised manner, specifically trained for the image patches present in each cluster. The experiments highlight the adaptability and efficacy through enhanced heterogeneous filtering, both from synthesized artifacts but also realistic SIDD image pairs.

eess.IV↗

arXiv2023

A Survey on Reinforcement Learning for Combinatorial Optimization

This paper gives a detailed review of reinforcement learning (RL) in combinatorial optimization, introduces the history of combinatorial optimization starting in the 1950s, and compares it with the RL algorithms of recent years. This paper explicitly looks at a famous combinatorial problem-traveling salesperson problem (TSP). It compares the approach of modern RL algorithms for the TSP with an approach published in the 1970s. By comparing the similarities and variances between these methodologies, the paper demonstrates how RL algorithms are optimized due to the evolution of machine learning techniques and computing power. The paper then briefly introduces the deep learning approach to the TSP named deep RL, which is an extension of the traditional mathematical framework. In deep RL, attention and feature encoding mechanisms are introduced to generate near-optimal solutions. The survey shows that integrating the deep learning mechanism, such as attention with RL, can effectively approximate the TSP. The paper also argues that deep learning could be a generic approach that can be integrated with any traditional RL algorithm to enhance the outcomes of the TSP.

cs.LG↗

arXiv2023

Specification-Driven Video Search via Foundation Models and Formal Verification

The increasing abundance of video data enables users to search for events of interest, e.g., emergency incidents. Meanwhile, it raises new concerns, such as the need for preserving privacy. Existing approaches to video search require either manual inspection or a deep learning model with massive training. We develop a method that uses recent advances in vision and language models, as well as formal methods, to search for events of interest in video clips automatically and efficiently. The method consists of an algorithm to map text-based event descriptions into linear temporal logic over finite traces (LTL$_f$) and an algorithm to construct an automaton encoding the video information. Then, the method formally verifies the automaton representing the video against the LTL$_f$ specifications and adds the pertinent video clips to the search result if the automaton satisfies the specifications. We provide qualitative and quantitative analysis to demonstrate the video-searching capability of the proposed method. It achieves over 90 percent precision in searching over privacy-sensitive videos and a state-of-the-art autonomous driving dataset.

cs.CV↗

arXiv2023

On the Planning, Search, and Memorization Capabilities of Large Language Models

The rapid advancement of large language models, such as the Generative Pre-trained Transformer (GPT) series, has had significant implications across various disciplines. In this study, we investigate the potential of the state-of-the-art large language model (GPT-4) for planning tasks. We explore its effectiveness in multiple planning subfields, highlighting both its strengths and limitations. Through a comprehensive examination, we identify areas where large language models excel in solving planning problems and reveal the constraints that limit their applicability. Our empirical analysis focuses on GPT-4's performance in planning domain extraction, graph search path planning, and adversarial planning. We then propose a way of fine-tuning a domain-specific large language model to improve its Chain of Thought (CoT) capabilities for the above-mentioned tasks. The results provide valuable insights into the potential applications of large language models in the planning domain and pave the way for future research to overcome their limitations and expand their capabilities.

cs.CL↗

arXiv2023

Automaton-Based Representations of Task Knowledge from Generative Language Models

Automaton-based representations of task knowledge play an important role in control and planning for sequential decision-making problems. However, obtaining the high-level task knowledge required to build such automata is often difficult. Meanwhile, large-scale generative language models (GLMs) can automatically generate relevant task knowledge. However, the textual outputs from GLMs cannot be formally verified or used for sequential decision-making. We propose a novel algorithm named GLM2FSA, which constructs a finite state automaton (FSA) encoding high-level task knowledge from a brief natural-language description of the task goal. GLM2FSA first sends queries to a GLM to extract task knowledge in textual form, and then it builds an FSA to represent this text-based knowledge. The proposed algorithm thus fills the gap between natural-language task descriptions and automaton-based representations, and the constructed FSA can be formally verified against user-defined specifications. We accordingly propose a method to iteratively refine the queries to the GLM based on the outcomes, e.g., counter-examples, from verification. We demonstrate GLM2FSA's ability to build and refine automaton-based representations of everyday tasks (e.g., crossing a road), and also of tasks that require highly-specialized knowledge (e.g., executing secure multi-party computation).

cs.FL↗

arXiv2023

Self-Enhancing Multi-filter Sequence-to-Sequence Model

Representation learning is important for solving sequence-to-sequence problems in natural language processing. Representation learning transforms raw data into vector-form representations while preserving their features. However, data with significantly different features leads to heterogeneity in their representations, which may increase the difficulty of convergence. We design a multi-filter encoder-decoder model to resolve the heterogeneity problem in sequence-to-sequence tasks. The multi-filter model divides the latent space into subspaces using a clustering algorithm and trains a set of decoders (filters) in which each decoder only concentrates on the features from its corresponding subspace. As for the main contribution, we design a self-enhancing mechanism that uses a reinforcement learning algorithm to optimize the clustering algorithm without additional training data. We run semantic parsing and machine translation experiments to indicate that the proposed model can outperform most benchmarks by at least 5\%. We also empirically show the self-enhancing mechanism can improve performance by over 10\% and provide evidence to demonstrate the positive correlation between the model's performance and the latent space clustering.

cs.CL↗

arXiv2022

Reinforcement Learning of Self Enhancing Camera Image and Signal Processing

Current camera image and signal processing pipelines (ISPs), including deep-trained versions, tend to apply a single filter that is uniformly applied to the entire image. This is despite the fact that most acquired camera images have spatially heterogeneous artifacts. This spatial heterogeneity manifests itself across the image space as varied Moire ringing, motion-blur, color-bleaching, or lens-based projection distortions. Moreover, combinations of these image artifacts can be present in small or large pixel neighborhoods, within an acquired image. Here, we present a deep reinforcement learning model that works in learned latent subspaces, and recursively improves camera image quality through a patch-based spatially adaptive artifact filtering and image enhancement. Our \textit{Recursive Self Enhancement Reinforcement Learning}(RSE-RL) model views the identification and correction of artifacts as a recursive self-learning and self-improvement exercise and consists of two major sub-modules: (i) The latent feature sub-space clustering/grouping obtained through variational auto-encoders enabling rapid identification of the correspondence and discrepancy between noisy and clean image patches. (ii) The adaptive learned transformation is controlled by a soft actor-critic agent that progressively filters and enhances the noisy patches using its closest feature distance neighbors of clean patches. Artificial artifacts that may be introduced in a patch-based ISP, are also removed through a reward-based de-blocking recovery and image enhancement. We demonstrate the self-improvement feature of our model by recursively training and testing on images, wherein the enhanced images resulting from each epoch provide a natural data augmentation and robustness to the RSE-RL training-filtering pipeline. Our method shows advantage for heterogeneous noise and artifact removal.

cs.CV↗

arXiv2022

On the Privacy Risks of Deploying Recurrent Neural Networks in Machine Learning Models

We study the privacy implications of training recurrent neural networks (RNNs) with sensitive training datasets. Considering membership inference attacks (MIAs), which aim to infer whether or not specific data records have been used in training a given machine learning model, we provide empirical evidence that a neural network's architecture impacts its vulnerability to MIAs. In particular, we demonstrate that RNNs are subject to a higher attack accuracy than feed-forward neural network (FFNN) counterparts. Additionally, we study the effectiveness of two prominent mitigation methods for preempting MIAs, namely weight regularization and differential privacy. For the former, we empirically demonstrate that RNNs may only benefit from weight regularization marginally as opposed to FFNNs. For the latter, we find that enforcing differential privacy through either of the following two methods leads to a less favorable privacy-utility trade-off in RNNs than alternative FFNNs: (i) adding Gaussian noise to the gradients calculated during training as a part of the so-called DP-SGD algorithm and (ii) adding Gaussian noise to the trainable parameters as a part of a post-training mechanism that we propose. As a result, RNNs can also be less amenable to mitigation methods, bringing us to the conclusion that the privacy risks pertaining to the recurrent architecture are higher than the feed-forward counterparts.

cs.CR↗

arXiv2022

Training Heterogeneous Features in Sequence to Sequence Tasks: Latent Enhanced Multi-filter Seq2Seq Model

In language processing, training data with extremely large variance may lead to difficulty in the language model's convergence. It is difficult for the network parameters to adapt sentences with largely varied semantics or grammatical structures. To resolve this problem, we introduce a model that concentrates the each of the heterogeneous features in the input sentences. Building upon the encoder-decoder architecture, we design a latent-enhanced multi-filter seq2seq model (LEMS) that analyzes the input representations by introducing a latent space transformation and clustering. The representations are extracted from the final hidden state of the encoder and lie in the latent space. A latent space transformation is applied for enhancing the quality of the representations. Thus the clustering algorithm can easily separate samples based on the features of these representations. Multiple filters are trained by the features from their corresponding clusters, and the heterogeneity of the training data can be resolved accordingly. We conduct two sets of comparative experiments on semantic parsing and machine translation, using the Geo-query dataset and Multi30k English-French to demonstrate the enhancement our model has made respectively.

cs.CL↗

arXiv2022

Additive Logistic Mechanism for Privacy-Preserving Self-Supervised Learning

We study the privacy risks that are associated with training a neural network's weights with self-supervised learning algorithms. Through empirical evidence, we show that the fine-tuning stage, in which the network weights are updated with an informative and often private dataset, is vulnerable to privacy attacks. To address the vulnerabilities, we design a post-training privacy-protection algorithm that adds noise to the fine-tuned weights and propose a novel differential privacy mechanism that samples noise from the logistic distribution. Compared to the two conventional additive noise mechanisms, namely the Laplace and the Gaussian mechanisms, the proposed mechanism uses a bell-shaped distribution that resembles the distribution of the Gaussian mechanism, and it satisfies pure $ε$-differential privacy similar to the Laplace mechanism. We apply membership inference attacks on both unprotected and protected models to quantify the trade-off between the models' privacy and performance. We show that the proposed protection algorithm can effectively reduce the attack accuracy to roughly 50\%-equivalent to random guessing-while maintaining a performance loss below 5\%.

cs.LG↗

arXiv2021

Identifying Mislabeled Images in Supervised Learning Utilizing Autoencoder

Supervised learning is based on the assumption that the ground truth in the training data is accurate. However, this may not be guaranteed in real-world settings. Inaccurate training data will result in some unexpected predictions. In image classification, incorrect labels may cause the classification model to be inaccurate as well. In this paper, I am going to apply unsupervised techniques to the training data before training the classification network. A convolutional autoencoder is applied to encode and reconstruct images. The encoder will project the image data on to latent space. In the latent space, image features are preserved in a lower dimension. The assumption is that data samples with similar features are likely to have the same label. Noised samples can be classified in the latent space by the Density-Base Scan (DBSCAN) clustering algorithm. These incorrectly labeled data are visualized as outliers in the latent space. Therefore, the outliers identified by the DBSCAN algorithm can be classified as incorrectly labeled samples. After the outliers are detected, all the outliers are treated as mislabeled data samples and removed from the dataset. Thus the training data can be directly used in training the supervised learning network. The algorithm can detect and remove above 67\% of mislabeled data in the experimental dataset.

cs.CV↗

arXiv2020

Approximating Boolean Functions with Disjunctive Normal Form

The theorem states that: Every Boolean function can be $ε-approximated$ by a Disjunctive Normal Form (DNF) of size $O_ε(2^{n}/\log{n})$. This paper will demonstrate this theorem in detail by showing how this theorem is generated and proving its correctness. We will also dive into some specific Boolean functions and explore how these Boolean functions can be approximated by a DNF whose size is within the universal bound $O_ε(2^{n}/\log{n})$. The Boolean functions we interested in are: Parity Function: the parity function can be $ε-approximated$ by a DNF of width $(1 - 2ε)n$ and size $2^{(1 - 2ε)n}$. Furthermore, we will explore the lower bounds on the DNF's size and width. Majority Function: for every constant $1/2 < ε< 1$, there is a DNF of size $2^{O(\sqrt{n})}$ that can $ε-approximated$ the Majority Function on n bits. Monotone Functions: every monotone function f can be $ε-approximated$ by a DNF g of size $2^{n - Ωε(n)}$ satisfying $g(x) \le f(x)$ for all x.

cs.CC↗