This research shows how to learn shared representations from unpaired data.
problem Learning shared representations from unpaired data.
method Spectral embeddings of random walk matrices from each unimodal representation.
result Shared representations can be learned almost exclusively from unpaired data.
Enhancing spectral embedding for low-dimensional embeddings in rare disease cohorts
problem Representing clinical concepts and patients in electronic health records
method Spectral-based unsupervised learning with flexible knowledge transfer
result Outperforms competing approaches in challenging scenarios
In deep neural nets, lower level embedding layers account for a large portion of the total number of parameters. Tikhonov regularization, graph-based regularization, and hard parameter sharing are approaches that introduce explicit biases into training in a hope to reduce statistical complexity. Alternatively, we propo…
The creation of social ties is largely determined by the entangled effects of people's similarities in terms of individual characters and friends. However, feature and structural characters of people usually appear to be correlated, making it difficult to determine which has greater responsibility in the formation of t…
Proposes EOT eigenmaps for aligning and embedding multiple datasets.
problem Aligning and embedding multiple datasets with shared structures but individual distortions.
method Entropic Optimal Transport (EOT) eigenmaps, leveraging leading singular vectors of EOT plan matrix.
result Proves theoretical guarantees and favorable properties for aligning and embedding datasets.
Recent works reveal that network embedding techniques enable many machine learning models to handle diverse downstream tasks on graph structured data. However, as previous methods usually focus on learning embeddings for a single network, they can not learn representations transferable on multiple networks. Hence, it i…
Kernel method embeds noisy datasets, capturing shared structures.
problem Limited power in capturing nonlinear structures, noisiness, high-dimensionality, and interpretability issues.
method Kernel spectral joint embeddings using duo-landmark integral operators.
result Consistent recovery of low-dimensional noiseless signals and convergence to eigenfunctions of integral operators.
Word embeddings are a powerful approach for analyzing language, and exponential family embeddings (EFE) extend them to other types of data. Here we develop structured exponential family embeddings (S-EFE), a method for discovering embeddings that vary across related groups of data. We study how the word usage of U.S. C…
Bayesian neural networks improve knowledge sharing across networks.
problem Improving knowledge sharing between neural networks.
method Application of Bayesian interpretations to neural networks for feature sharing.
result Enhanced feature sharing across networks.
GDA-HIN adapts across heterogeneous networks by aligning shared and private node types.
problem Domain adaptation challenges in heterogeneous networks with shared and private node types.
method Generalized Domain Adaptive model across HINs (GDA-HIN) that aligns identical-type nodes and edges while utilizing different-type nodes and edges.
result GDA-HIN outperforms state-of-the-art methods in various domain adaptation tasks across heterogeneous networks.
Despite significant progress, deep reinforcement learning (RL) suffers from data-inefficiency and limited generalization. Recent efforts apply meta-learning to learn a meta-learner from a set of RL tasks such that a novel but related task could be solved quickly. Though specific in some ways, different tasks in meta-RL…
ALF reduces network parameters and operations by 70% and 61%, respectively, on embedded hardware.
problem Efficient deployment of deep learning models on resource-constrained hardware.
method Autoencoder-based low-rank filter-sharing technique.
result ALF achieves significant compression with minimal accuracy loss.
Simple framework decouples word alignment and multilingual embedding mapping.
problem Learning multilingual embeddings without supervision.
method Two-stage approach: 1) unsupervised word alignment, 2) mapping embeddings to shared space.
result Robust performance across various multilingual tasks, including distant languages.
DECAT framework evaluates multimodal models for shared biology, detecting confounders and false positives.
problem Determining if multimodal models learn shared biology or just confounders.
method DECAT framework classifies multimodal representations into four diagnostic scenarios using null-referenced metrics.
result DECAT detects confounders and false positives in multimodal models, improving with larger cohorts and stronger representations.
Acoustic word embeddings --- fixed-dimensional vector representations of arbitrary-length words --- have attracted increasing interest in query-by-example spoken term detection. Recently, on the fact that the orthography of text labels partly reflects the phonetic similarity between the words' pronunciation, a multi-vi…
Every day, hundreds of millions of new Tweets containing over 40 languages of ever-shifting vernacular flow through Twitter. Models that attempt to extract insight from this firehose of information must face the torrential covariate shift that is endemic to the Twitter platform. While regularly-retrained algorithms can…
Method embeds numeric tabular datasets into a shared vector space for similarity and retrieval.
problem Lack of meaningful representation for numeric tabular datasets in large language models.
method Structured exploratory data analysis descriptors, sentence transformer embedding, CCA for cross-dataset alignment.
result Total P@1 score of 0.9 across 15 datasets, robust nearest-neighbor retrieval and cluster structure.
Paper proposes a new topology for AML analysis using Poincaré embeddings.
problem Complex money laundering schemes and regulatory constraints hinder AML analysis and information sharing.
method Proposes a new topology for AML analysis using Poincaré embeddings.
result Demonstrates improved AML analysis and information sharing through Poincaré embeddings.
Anchor PCA improves robustness in multi-domain PCA.
problem PCA on pooled data can focus on spurious directions.
method Anchor PCA focuses on shared directions of variation.
result Anchor PCA outperforms pooling and worst-case alternatives.
SPIRE enables efficient federated learning for diffusion models by separating client-specific embeddings from a shared backbone.
problem Large diffusion models are impractical for federated learning due to their size.
method SPIRE separates the network into a global backbone and client-specific embeddings, enabling efficient finetuning.
result SPIRE achieves parameter-efficient finetuning, updating only a small fraction of weights.
CMF is a technique for simultaneously learning low-rank representations based on a collection of matrices with shared entities. A typical example is the joint modeling of user-item, item-property, and user-feature matrices in a recommender system. The key idea in CMF is that the embeddings are shared across the matrice…
MCPCA analyzes shared factors across multiple data contexts.
problem No tools to recover shared factors across multiple contexts.
method Developed a theoretical and algorithmic framework (MCPCA).
result Reveals shared axes of variation across subsets of contexts.
Graph embedding leaks sensitive graph properties and subgraphs.
problem Privacy risks in graph embedding sharing.
method Three inference attacks and a defense mechanism.
result High accuracy in inferring graph properties and subgraphs.
MPP trains a transformer to predict multiple physical systems, improving accuracy across various tasks.
problem Training models for specific physical systems is inefficient and requires fine-tuning.
method MPP trains a shared transformer on multiple heterogeneous physical systems, projecting fields into a shared embedding space.
result A single MPP-pretrained transformer outperforms task-specific models on all pretraining sub-tasks and downstream tasks.
Enhances generative model accuracy through knowledge transfer.
problem Improving generative model precision across different tasks.
method Introduces a novel framework for transfer learning using shared structures.
result Demonstrates enhanced performance in diffusion and normalizing flows models.
A new method combines multiple node embeddings using tensor decomposition.
problem Generating accurate node embeddings for complex networks.
method TenSemble2Vec: combines multiple embeddings via tensor decomposition.
result Improves node embeddings by leveraging complementary information from different methods.
Method captures shared information across many views robustly.
problem Modeling hundreds of views per event and learning robust embeddings without view knowledge.
method View bootstrapping using multi-view correlation and matrix concentration theory.
result View bootstrapping captures shared information across many views robustly.
We propose a novel framework for multi-task reinforcement learning (MTRL). Using a variational inference formulation, we learn policies that generalize across both changing dynamics and goals. The resulting policies are parametrized by shared parameters that allow for transfer between different dynamics and goal condit…
Enhances graph classification with multiple graphs.
problem Improving graph classification accuracy with multiple graphs.
method Graph fusion embedding using encoder embedding.
result The method consistently improves classification accuracy for large vertex sets.
New approach for sharing deep learning costs between devices and cloud.
problem Prohibitive deep learning computational requirements for embedded devices.
method Study of representation compressibility in MobileNetV2 for balancing computation, bandwidth, and accuracy.
result An optimal splitting layer for network can be found with a simple PCA-based compression scheme.
A method for learning embeddings from multi-view data using Gromov-Wasserstein.
problem Challenges in learning low-dimensional representations from multi-view relational data with differing geometries.
method Bary-GWMDS and Mean-GWMDS-C, Gromov-Wasserstein-based methods operating on distance matrices.
result Stable and geometrically meaningful embeddings learned from synthetic and real-world datasets.
This work explains how maximizing latent correlations across multiple data views helps in identifying shared and private components.
problem Understanding how to identify shared and private components in multiview data.
method An intuitive generative model of multiview data is adopted, and latent correlation maximization is shown to guarantee the extraction of shared components.
result Latent correlation maximization guarantees the extraction of shared components across views and disentangles private information.
Proposes MANE for multi-view network embedding, improving node representations.
problem Learning low-dimensional representations from multiple views of networks.
method MANE combines diversity and collaboration, including second-order collaboration, and attention-based extension MANE+.
result MANE+ outperforms state-of-the-art approaches on real-world multi-view networks.
The paper constructs λ-hypersurfaces for λ>0 and λ<0.
problem Exploring λ-hypersurfaces in different λ-values and their properties. method Constructing complete embedded and non-convex λ-hypersurfaces diffeomorphic to a cylinder and doughnut-shaped. result For λ>0, complete embedded and non-convex λ-hypersurfaces are constructed, diffeomorphic to a cylinder. The hyperbolic manifold is a smooth manifold of negative constant curvature. While the hyperbolic manifold is well-studied in the literature, it has gained interest in the machine learning and natural language processing communities lately due to its usefulness in modeling continuous hierarchies. Tasks with hierarchica…
Study uses trajectory embedding to measure place function similarity at fine spatial granularity.
problem Measuring place function similarity at fine spatial granularity.
method Trajectory embedding to reduce dimensions and measure similarity of place functions.
result Embedding similarity can be a metric proxy for place functions at fine spatial granularity.
PIMA autoencoders discover shared features in multimodal scientific data.
problem Discovering shared information in high-throughput scientific datasets.
method Physics-informed multimodal autoencoders (PIMA) with Gaussian mixture prior and product of experts formulation.
result Accurate cross-modal inference between images and mechanical stress-strain response in lattice metamaterials.
The heterogeneity-gap between different modalities brings a significant challenge to multimedia information retrieval. Some studies formalize the cross-modal retrieval tasks as a ranking problem and learn a shared multi-modal embedding space to measure the cross-modality similarity. However, previous methods often esta…
Representation learning has recently been successfully used to create vector representations of entities in language learning, recommender systems and in similarity learning. Graph embeddings exploit the locality structure of a graph and generate embeddings for nodes which could be words in a language, products of a re…
A new method is proposed to obtain the risk neutral probability of share prices without stochastic calculus and price modeling, via an embedding of the price return modeling problem in Le Cam's statistical experiments framework. Strategies-probabilities Pt0,n and PT,n are thus determined and used, respective…
For embedded 2-spheres in a 4-manifold sharing the same embedded transverse sphere homotopy implies isotopy, provided the ambient 4-manifold has no $\BZ_2$-torsion in the fundamental group. This gives a generalization of the classical light bulb trick to 4-dimensions, the uniqueness of spanning discs for a simple close…
Prototypical Networks improve multi-label classification accuracy.
problem Multi-label classification with nonlinear label dependencies.
method Formulate multi-label learning as class distribution in a non-linear embedding space. For each label, positive and negative embeddings are compactly distributed. Labels are inferred by measuring the distance to prototype positive or negative embeddings.
result Extensive experiments show improved accuracy compared to state-of-the-art algorithms.
Training-free source selection for LLM families with shared vocabularies
problem Source selection for LLM families with shared vocabularies
method Fisher alignment at vocabulary scale
result Fisher alignment is a cosine between kernel mean embeddings in the joint activation-error space
Traditionally, many text-mining tasks treat individual word-tokens as the finest meaningful semantic granularity. However, in many languages and specialized corpora, words are composed by concatenating semantically meaningful subword structures. Word-level analysis cannot leverage the semantic information present in su…
This paper studies how to find compact state embeddings from high-dimensional Markov state trajectories, where the transition kernel has a small intrinsic rank. In the spirit of diffusion map, we propose an efficient method for learning a low-dimensional state embedding and capturing the process's dynamics. This idea a…
Word embedding maps words into a low-dimensional continuous embedding space by exploiting the local word collocation patterns in a small context window. On the other hand, topic modeling maps documents onto a low-dimensional topic space, by utilizing the global word collocation patterns in the same document. These two …
The paper describes complex structures on Oeljeklaus-Toma manifolds.
problem Investigating different complex structures on Oeljeklaus-Toma manifolds.
method Algebraic description of biholomorphic manifolds obtained by conjugating complex embeddings.
result Compact non-Kähler manifolds with multiple rigid complex structures.
IITK wins FinSim 2020 task on financial hypernym detection.
problem Classifying financial terms into hypernym concepts in an external ontology.
method Used context-dependent and context-independent word embeddings (Word2vec and BERT) for classification.
result Ranked 1st based on mean rank and accuracy metrics.