Physical Sciences and Mathematics Commons^™

Identifiability Of Label Noise Transition Matrix, Yang Liu, Hao Cheng, Kun Zhang

Machine Learning Faculty Publications

The noise transition matrix plays a central role in the problem of learning with noisy labels. Among many other reasons, a large number of existing solutions rely on the knowledge of it. Identifying and estimating the transition matrix without ground truth labels is a critical and challenging task. When label noise transition depends on each instance, the problem of identifying the instance-dependent noise transition matrix becomes substantially more challenging. Despite recently proposed solutions for learning from instance-dependent noisy labels, the literature lacks a unified understanding of when such a problem remains identifiable. The goal of this paper is to characterize …

Adaptive Compositional Continual Meta-Learning, Bin Wu, Jinyuan Fang, Xiangxiang Zeng, Shangsong Liang, Qiang Zhang

Machine Learning Faculty Publications

This paper focuses on continual meta-learning, where few-shot tasks are heterogeneous and sequentially available. Recent works use a mixture model for meta-knowledge to deal with the heterogeneity. However, these methods suffer from parameter inefficiency caused by two reasons: (1) the underlying assumption of mutual exclusiveness among mixture components hinders sharing meta-knowledge across heterogeneous tasks. (2) they only allow increasing mixture components and cannot adaptively filter out redundant components. In this paper, we propose an Adaptive Compositional Continual Meta-Learning (ACML) algorithm, which employs a compositional premise to associate a task with a subset of mixture components, allowing meta-knowledge sharing among heterogeneous …

Fast Rates For Maximum Entropy Exploration, Daniil Tiapkin, Denis Belomestny, Daniele Calandriello, Éric Moulines, Rémi Munos, Alexey Naumov, Pierre Perrault, Yunhao Tang, Michal Valko, Pierre Ménard

Machine Learning Faculty Publications

We address the challenge of exploration in reinforcement learning (RL) when the agent operates in an unknown environment with sparse or no rewards. In this work, we study the maximum entropy exploration problem of two different types. The first type is visitation entropy maximization previously considered by Hazan et al. (2019) in the discounted setting. For this type of exploration, we propose a game-theoretic algorithm that has Oe(H3S2A/ε2) sample complexity thus improving the ε-dependence upon existing results, where S is a number of states, A is a number of actions, H is an episode length, and ε is a desired …

Evolving Semantic Prototype Improves Generative Zero-Shot Learning, Shiming Chen, Wenjin Hou, Ziming Hong, Xiaohan Ding, Yibing Song, Xinge You, Tongliang Liu, Kun Zhang

Machine Learning Faculty Publications

In zero-shot learning (ZSL), generative methods synthesize class-related sample features based on predefined semantic prototypes. They advance the ZSL performance by synthesizing unseen class sample features for better training the classifier. We observe that each class's predefined semantic prototype (also referred to as semantic embedding or condition) does not accurately match its real semantic prototype. So the synthesized visual sample features do not faithfully represent the real sample features, limiting the classifier training and existing ZSL performance. In this paper, we formulate this mismatch phenomenon as the visual-semantic domain shift problem. We propose a dynamic semantic prototype evolving (DSP) method …

Coco: A Coupled Contrastive Framework For Unsupervised Domain Adaptive Graph Classification, Nan Yin, Li Shen, Mengzhu Wang, Long Lan, Zeyu Ma, Chong Chen, Xian Sheng Hua, Xiao Luo

Machine Learning Faculty Publications

Although graph neural networks (GNNs) have achieved impressive achievements in graph classification, they often need abundant task-specific labels, which could be extensively costly to acquire. A credible solution is to explore additional labeled graphs to enhance unsupervised learning on the target domain. However, how to apply GNNs to domain adaptation remains unsolved owing to the insufficient exploration of graph topology and the significant domain discrepancy. In this paper, we propose Coupled Contrastive Graph Representation Learning (CoCo), which extracts the topological information from coupled learning branches and reduces the domain discrepancy with coupled contrastive learning. CoCo contains a graph convolutional network …

Convergence Of Proximal Point And Extragradient-Based Methods Beyond Monotonicity: The Case Of Negative Comonotonicity, Eduard Gorbunov, Adrien Taylor, Samuel Horváth, Gauthier Gidel

Machine Learning Faculty Publications

Algorithms for min-max optimization and variational inequalities are often studied under monotonicity assumptions. Motivated by non-monotone machine learning applications, we follow the line of works (Diakonikolas et al., 2021; Lee & Kim, 2021; Pethick et al., 2022; Böhm, 2022) aiming at going beyond monotonicity by considering the weaker negative comonotonicity assumption. In this work, we provide tight complexity analyses for the Proximal Point (PP), Extragradient (EG), and Optimistic Gradient (OG) methods in this setup, closing several questions on their working guarantees beyond monotonicity. In particular, we derive the first non-asymptotic convergence rates for PP under negative comonotonicity and star-negative comonotonicity …

Unleashing Mask: Explore The Intrinsic Out-Of-Distribution Detection Capability, Jianing Zhu, Hengzhuang Li, Jiangchao Yao, Tongliang Liu, Jianliang Xu, Bo Han

Machine Learning Faculty Publications

Out-of-distribution (OOD) detection is an indispensable aspect of secure AI when deploying machine learning models in real-world applications. Previous paradigms either explore better scoring functions or utilize the knowledge of outliers to equip the models with the ability of OOD detection. However, few of them pay attention to the intrinsic OOD detection capability of the given model. In this work, we generally discover the existence of an intermediate stage of a model trained on in-distribution (ID) data having higher OOD detection performance than that of its final stage across different settings, and further identify one critical data-level attribution to be …

Exploring Model Dynamics For Accumulative Poisoning Discovery, Jianing Zhu, Xiawei Guo, Jiangchao Yao, Chao Du, Li He, Shuo Yuan, Tongliang Liu, Liang Wang, Bo Han

Machine Learning Faculty Publications

Adversarial poisoning attacks pose huge threats to various machine learning applications. Especially, the recent accumulative poisoning attacks show that it is possible to achieve irreparable harm on models via a sequence of imperceptible attacks followed by a trigger batch. Due to the limited data-level discrepancy in real-time data streaming, current defensive methods are indiscriminate in handling the poison and clean samples. In this paper, we dive into the perspective of model dynamics and propose a novel information measure, namely, Memorization Discrepancy, to explore the defense via the model-level information. By implicitly transferring the changes in the data manipulation to that …

Causal Discovery With Latent Confounders Based On Higher-Order Cumulants, Ruichu Cai, Zhiyi Huang, Wei Chen, Zhifeng Hao, Kun Zhang

Machine Learning Faculty Publications

Causal discovery with latent confounders is an important but challenging task in many scientific areas. Despite the success of some overcomplete independent component analysis (OICA) based methods in certain domains, they are computationally expensive and can easily get stuck into local optima. We notice that interestingly, by making use of higher-order cumulants, there exists a closed-form solution to OICA in specific cases, e.g., when the mixing procedure follows the One-Latent-Component structure. In light of the power of the closed-form solution to OICA corresponding to the One-Latent-Component structure, we formulate a way to estimate the mixing matrix using the higher-order cumulants, …

Feature Expansion For Graph Neural Networks, Jiaqi Sun, Lin Zhang, Guangyi Chen, Peng Xu, Kun Zhang, Yujiu Yang

Machine Learning Faculty Publications

Graph neural networks aim to learn representations for graph-structured data and show impressive performance, particularly in node classification. Recently, many methods have studied the representations of GNNs from the perspective of optimization goals and spectral graph theory. However, the feature space that dominates representation learning has not been systematically studied in graph neural networks. In this paper, we propose to fill this gap by analyzing the feature space of both spatial and spectral models. We decompose graph neural networks into determined feature spaces and trainable weights, providing the convenience of studying the feature space explicitly using matrix space analysis. In …

A Unified Optimization Framework Of Ann-Snn Conversion: Towards Optimal Mapping From Activation Values To Firing Rates, Haiyan Jiang, Srinivas Anumasa, Giulia De Masi, Huan Xiong, Bin Gu

Machine Learning Faculty Publications

Spiking Neural Networks (SNNs) have gained significant attention for their energy-efficient and fast-inference capabilities, but training SNNs from scratch can be challenging due to the discrete nature of spikes. One alternative method is to convert an Artificial Neural Network (ANN) into an SNN, known as ANN-SNN conversion. Currently, existing ANN-SNN conversion methods often involve redesigning the ANN with a new activation function, rather than utilizing the traditional ReLU, and converting it to an SNN. However, these methods do not take into account the potential performance loss between the regular ANN with ReLU and the tailored ANN. In this work, we …

Mentoring Experiences Of Undergraduate Students And Faculty Members In Science, Technology, Engineering, And Mathematics, Pamela Martínez Oquendo

School of Natural Resources: Dissertations, Theses, and Student Research

I present a comprehensive view of mentoring experiences of undergraduate students and faculty members in science, technology, engineering, and mathematics (STEM). In CHAPTER 1, I describe a brief outline of this dissertation. In CHAPTER 2, I present an interpretative phenomenological analysis of the lived experiences of former STEM undergraduate mentors of the Nebraska STEM For You (NE STEM 4U) afterschool mentoring program. In CHAPTER 3, I describe how the ramifications of faculty mentorship influence the science pipeline using a qualitative synthesis. In CHAPTER 4, I describe how the STEM faculty-student mentoring engagement involves a strong psychological support component using a …

Current Vehicle Fleet Inventory And Future Implementation Of A Centralized Electric Fleet At Portland State University, Dane Kovaleski

Environmental Science and Management Professional Master's Project Reports

As the effects of climate change continue to impact the world, many institutions have developed climate action goals to reduce their effects on the environment. Portland State University (PSU) has committed to an 80% reduction of greenhouse gas emissions by 2030 and carbon neutrality by 2040. A part of this commitment must include looking at the contributions of transportation on campus to reduce carbon emissions. According to a greenhouse gas emissions report done by the Campus Planning and Sustainability Office in 2016, transportation contributed to 12% of total greenhouse gas emissions on campus.

This project aims to evaluate the management …

Computational Model Of Twisted Elastic Ribbons, Madelyn Leembruggen, Jovana Andrejevic, Arshad Kudrolli, Chris H. Rycroft

Physics

We develop an irregular lattice mass-spring model to simulate and study the deformation modes of a thin elastic ribbon as a function of applied end-to-end twist and tension. Our simulations reproduce all reported experimentally observed modes, including transitions from helicoids to longitudinal wrinkles, creased helicoids and loops with self-contact, and transverse wrinkles to accordion self-folds. Our simulations also show that the twist angles at which the primary longitudinal and transverse wrinkles appear are well described by various analyses of the Föppl-von Kármán equations, but the characteristic wavelength of the longitudinal wrinkles has a more complex relationship to applied tension than …

Density-Mediated Spin Correlations Drive Edge-To-Bulk Flow Transition In Active Chiral Matter, Alexander P. Petroff, Christopher Whittington, Arshad Kudrolli

Physics

We demonstrate that edge currents develop in active chiral matter due to boundary shielding over a wide range of densities corresponding to a gas, fluid, and crystal. The system is composed of spinning disk-shaped grains with chirally arranged tilted legs confined in a circular vibrating chamber. The edge currents are shown to increasingly drive circulating bulk flows with area fraction as percolating clusters develop due to increasing spin-coupling between neighbors mediated by frictional contacts. Edge currents are observed even in the dilute limit. While, at low area fraction, the average flux vanishes except within a distance that is of the …

Long Island Sound Reef Gis Data, Robert M. Cerrato, Matthew Sclafani

SoMAS Research Data

High-resolution backscatter and bathymetric maps created by multibeam sonar surveys were used to identify different seafloor bottom types within existing, potentially expanded, and newly proposed reef areas in New York waters. Existing sites included Smithtown in Long Island Sound (LIS), and Rockaway, Atlantic Beach, Hempstead, Yellowbar, Kismet, Fire Island, Twelve Mile along the South Shore. Potential expansions are proposed on the South Shore for McAllister, Moriches, and Shinnecock reefs in addition to a new site called Sixteen Fathom. In Long Island Sound, new sites are proposed for Huntington/Oyster Bay, Port Jefferson/Mount Sinai, and Mattituck. Grab samples were collected within these …

Numerical Design And Optimization Of Near-Infrared Band- Pass Filter, Hafiza Syeeda Faiza, Ghazi Aman Nowsherwan, Basem A. Abu Izneid, Muhammad Azhar, Saira Riaz, Syed Sajjad Hussain, Saira Ikram, Mohsin Khan, Shahzad Naseem, Mohammad Kanan, Ibrahim M. Mansour

Applied Mathematics & Information Sciences

Band-pass filters functioning in the near-infrared (IR) range are desired for laser technology, multi-photon fluorescence, and IR imaging applications. In this study, we have designed four band-pass filters in the near Infrared spectrum (900-1200 nm) by vertically stacking different high and low-index materials. The band-pass filters are modelled by Essential Macleod software with different thicknesses. The layer’s thicknesses were optimized in such a way to provide the negligible reflectance and maximum transmission on the front side. All the simulated band-pass filters exhibit high transmittance, but TiO2/Al2O3 and Ta2O5/Al2O3 outperforms other modelled structure in terms of performance due to the better …

Extending The Convolution In Graph Neural Networks To Solve Materials Science And Node Classification Problems, Steph-Yves Mike Louis

Theses and Dissertations

The usage of graph to represent one's data in machine learning has grown in popularity in both academia and the industry due to its inherent benefits. With its flexible nature and immediate translation to real life observed objects, graph representation had a considerable contribution in advancing the state-of-the-art performance of machine learning in materials.

In this dissertation proposal, we discuss how machines can learn from graph encoded data and provide excellent results through graph neural networks (GNN). Notably, we focus our adaptation of graph neural networks on three tasks: predicting crystal materials properties, nullifying the negative impact of inferior graph …

Humans Against Large Language Models On Hard Paraphrase Detection Tasks, Jamie C. Macbeth, Ella Chang, Jingyu Gin Chen, Yining Hua, Sandra Grandic, Winnie X. Zheng

Computer Science: Faculty Publications

The ability to recognize that pairs or sets of language expressions “mean the same thing” is a cognitive task for which meaning representation is clearly a central issue. This paper uses the task of paraphrasing to study meaning representation in a cognitive system. The main claim is that a consequential part of the meaning representation for a natural language expression is a set of language-free structures that are not part of the expression in question. To support this claim, we construct a corpus of paraphrase pairs using a system that has a non-linguistic meaning represen- tation decoupled from the linguistic …

Dynamic Fallowing In The Middle Rio Grande: A Look At The Environmental Water Leasing Program, Jared Wood

Water Resources Professional Project Reports

The Environmental Water Leasing Program (EWLP) is a federally-funded collaborative effort of the Middle Rio Grande Conservancy District and the U.S. Bureau of Reclamation that has paid irrigators in the Middle Rio Grande to temporarily lease water rights and forgo their water use so that it may be used to provide environmental flows for the endangered Rio Grande Silvery Minnow. The program operates in a sometimes socially contentious context, in a region with an overallocated, legally constrained water system, to offer much needed water management flexibility. While dynamic fallowing payments for ecosystem services programs like the EWLP are common policy …

Irrigation And Innovation: Understanding Barriers To Innovative Actions To Manage Drought On Middle Rio Grande Farms, Eleanor C. Hasenbeck

Water Resources Professional Project Reports

The Rio Grande has irrigated agricultural land in central New Mexico since time immemorial. Today, agriculture in New Mexico’s Middle Rio Grande valley is adapting to an increasingly scarce water supply due to long-term drought and climate change (Dunbar et al., 2022). As supply is decreasing, demands are increasing because of several factors: an increase in evapotranspiration (Dunbar et al., 2022); population growth in the Middle Rio Grande valley (NM LFC, 2021); and legal obligations to provide water for endangered species and to Texas and Mexico (NM ISC, 2017). Agriculture is largest water use by sector in the Rio Grande …

Logarithmic Terms In Atom-Surface Potentials: Limited Applicability Of Rational Approximations For Intermediate Distance, Ulrich D. Jentschura, C. Moore

Physics Faculty Research & Creative Works

It is usually assumed that interaction potentials, in general, and atom-surface potential, in particular, can be expressed in terms of an expansion involving integer powers of the distance between the two interacting objects. Here, we show that, in the short-range expansion of the interaction potential of a neutral atom and a dielectric surface, logarithms of the atom-wall distance appear. These logarithms are accompanied with logarithmic sums over virtual excitations of the atom interacting with the surface in analogy to Bethe logarithms in quantum electrodynamics. We verify the presence of the logarithmic terms in the short-range expansion using a model problem …

Galaxy Clustering In The Mira-Titan Universe. I. Emulators For The Redshift Space Galaxy Correlation Function And Galaxy-Galaxy Lensing, Juliana Kwan, Shun Saito, Alexie Leauthaud, Katrin Heitmann, Salman Habib, Nicholas Frontiere, Hong Guo, Song Huang, Adrian Pope, Sergio Rodriguéz-Torres

Physics Faculty Research & Creative Works

We construct accurate emulators for the projected and redshift space galaxy correlation functions and excess surface density as measured by galaxy-galaxy lensing, based on halo occupation distribution modeling. Using the complete Mira-Titan suite of 111 N-body simulations, our emulators vary over eight cosmological parameters and include the effects of neutrino mass and dynamical dark energy. We demonstrate that our emulators are sufficiently accurate for the analysis of the Baryon Oscillation Spectroscopic Survey DR12 CMASS galaxy sample over the range 0.5 ≤ r ≤ 50 h −1 Mpc. Furthermore, we show that our emulators are capable of recovering unbiased cosmological constraints …

Effects Of Landslides On Terrestrial Carbon Stocks With A Coupled Geomorphic-Biologic Model: Southeast Alaska, United States, Adam M. Booth, Brian Buma, S. Nagorski

Geology Faculty Publications and Presentations

Landslides influence the global carbon (C) cycle by facilitating transfer of terrestrial C in biomass and soils to offshore depocenters and redistributing C within the landscape, affecting the terrestrial C reservoir itself. How landslides affect terrestrial C stocks is rarely quantified, so we derive a model that couples stochastic landslides with terrestrial C dynamics, calibrated to temperate rainforests in southeast Alaska, United States. Modeled landslides episodically transfer C from scars to deposits and destroy living biomass. After a landslide, total C stocks on the scar recover, while those on the deposit either increase (in the case of living biomass) or …

Groundwater Budgeting And Climate Change Vulnerability Assessment Of Water Supply At Bosque Del Apache National Wildlife Refuge, Maximiliano Trujillo

Water Resources Professional Project Reports

Water resources management is becoming increasingly complex in the Middle Rio Grande Valley, where the negative impacts of climate change and growing demand for a fully allocated river put further strain on diminishing water supply. The precarity of future water availability emphasizes the importance of fully accounting for existing water rights allocations and understanding how ever changing social and ecological processes might hinder those rights from being fully realized. This study aimed at improving knowledge of the groundwater budget at Bosque del Apache National Wildlife Refuge, New Mexico by utilizing experimental methods of Darcy Flow analysis to estimate groundwater flow …

Motivations And Barriers To Participation In Community Outreach And Engagement Among Environmental And Water Resources Students And Postdocs, Sydney Donohue

Water Resources Professional Project Reports

There is a need to utilize knowledge exchange avenues between academia and non-academic communities in order to advance environmental science and engineering education. Many universities and research centers attempt to enhance knowledge sharing by organizing broader impact outreach events such as lab tours, demonstrations, hands-on activities, and public presentations. Anecdotally, at two NSF-funded centers, the Transformation Network and Center for Water and the Environment, we have seen that only the same small subset of students and postdocs frequently participate while over 70% of those invited never volunteer. This study aims to survey environmentally-focused undergraduate students, graduate students, and postdocs’ motivations …

An Ml Based Digital Forensics Software For Triage Analysis Through Face Recognition, Gaurav Gogia, Parag H. Rughani

Journal of Digital Forensics, Security and Law

Since the past few years, the complexity and heterogeneity of digital crimes has increased exponentially, which has made the digital evidence & digital forensics paramount for both criminal investigation and civil litigation cases. Some of the routine digital forensic analysis tasks are cumbersome and can increase the number of pending cases especially when there is a shortage of domain experts. While the work is not very complex, the sheer scale can be taxing. With the current scenarios and future predictions, crimes are only going to become more complex and the precedent of collecting and examining digital evidence is only going …

Pursue Sustainability, Stay Paranoid In A Post-Covid World, Ho Kwon Ping, Havovi Joshi

Asian Management Insights

Ho Kwon Ping, Founder and Executive Chairman of Banyan Tree Holdings, speaks to Havovi Joshi about making sure sustainability is more than just a buzzword, his optimism regarding Asia’s growth in the future, and the need for youths to think differently about their careers.

Team Thesyllogist At Semeval-2023 Task 3: Language-Agnostic Framing Detection In Multi-Lingual Online News: A Zero-Shot Transfer Approach, Osama Mohammed Afzal, Preslav Nakov

Natural Language Processing Faculty Publications

We describe our system for SemEval-2022 Task 3 subtask 2 which on detecting the frames used in a news article in a multi-lingual setup. We propose a multi-lingual approach based on machine translation of the input, followed by an English prediction model. Our system demonstrated good zero-shot transfer capability, achieving micro-F1 scores of 53% for Greek (4th on the leaderboard) and 56.1% for Georgian (3rd on the leaderboard), without any prior training on translated data for these languages. Moreover, our system achieved comparable performance on seven other languages, including German, English, French, Russian, Italian, Polish, and Spanish. Our results demonstrate …

Semeval-2023 Task 3: Detecting The Category, The Framing, And The Persuasion Techniques In Online News In A Multi-Lingual Setup, Jakub Piskorski, Nicolas Stefanovitch, Giovanni Da San Martino, Preslav Nakov

Natural Language Processing Faculty Publications

We describe SemEval-2023 task 3 on Detecting the Category, the Framing, and the Persuasion Techniques in Online News in a Multilingual Setup: the dataset, the task organization process, the evaluation setup, the results, and the participating systems. The task focused on news articles in nine languages (six known to the participants upfront: English, French, German, Italian, Polish, and Russian), and three additional ones revealed to the participants at the testing phase: Spanish, Greek, and Georgian). The task featured three subtasks: (1) determining the genre of the article (opinion, reporting, or satire), (2) identifying one or more frames used in an …