The effectiveness of advertising in e-commerce largely depends on the ability of merchants to bid on and win impressions for their targeted users. The bidding procedure is highly complex due to various factors such as market competition, user behavior, and the diverse objectives of advertisers. In this paper we consider the problem at the level of user timelines instead of individual bid requests, manipulating full policies (i.e. pre-defined bidding strategies) and not bid values. In order to optimally allocate policies to users, typical multiple treatments allocation methods solve knapsack-like problems which aim at maximizing an expected value under constraints. In the specific context of online advertising, we argue that optimizing for the probability of success is a more suited objective than expected value maximization, and we introduce the SuccessProbaMax algorithm that aims at finding the policy allocation which is the most likely to outperform a fixed reference policy. Finally, we conduct comprehensive experiments both on synthetic and real-world data to evaluate its performance. The results demonstrate that our proposed algorithm outperforms conventional expected-value maximization algorithms in terms of success rate.

Maximizing the Success Probability of Policy Allocations in Online Systems

Obstructive sleep apnea (OSA) affects millions of U.S. Veterans and is linked to severe illnesses such as stroke and heart failure. However, understanding how OSA interacts with common comorbidities and the additive risk of OSA in developing severe illnesses is lacking. Large language models (LLMs) are promising tools for clinical natural language processing (NLP). These models are pre-trained on vast amounts of text data and learn to generate coherent and contextually appropriate text by predicting tokens in a sequence. This study explores the quality of LLM embeddings in clustering disease phenotypes related to OSA and its comorbidities such as atrial fibrillation, coronary artery disease, heart failure, hypertension, stroke, and type 2 diabetes mellitus. By identifying sub-cohorts of patients with similar diagnoses, procedures, medications, and adherence to continuous positive airway pressure (CPAP) treatment, we aim to score LLM embeddings by how well pre-trained LLMs discriminate OSA-related phenotypes. Document-level embeddings for discharge reports in the MIMIC-IV database were computed using NERSC's Perlmutter supercomputer. Twelve clinical LLMs of different sizes and pre-training schemes were evaluated including BioBART, BioBERT, BioGPT, BioMegatron, Gatortron, and RadBERT. To visualize the LLMs embedding, UMAP was used to enable plotting of embedded documents in 2D. Each set was clustered using K-Means with increasing values for K, and the purity of the clusters was measured using the Shannon entropy of cluster assignments. Finally, we test the performance of quantized document-level data for supervised phenotype classification.
Qualitatively, we examined the LLM embeddings along UMAP plots and reviewed K-Means cluster assignment by reading sampled text. Gatrotron_medium, BioGPT_large, and BioBERT_base were the best models at separating OSA from its comorbidities. Quantitatively, the clusters were measured on their Shannon entropy factor in which the lower value correlated to greater purity of clusters and vice versa. Gatortron_medium, Gatortron_base, and BioBART_base consistently scored with low entropy across comorbidities. The developed method selected the Gatortron_medium LLM as the best pre-trained model for OSA-related phenotypes both qualitatively and quantitatively. The method also has potential for assessing bias in LLMs by measuring how embeddings group patients according to social determinants of health such as gender and race. In the context of the Department of Veterans Affairs (VA) healthcare, there is a need for early OSA detection and prevention. Leveraging LLMs for feature discovery and risk factor identification, this research can provide valuable insights for allocating CPAP treatment effectively by grouping patients with similar health histories and OSA outcomes.

Utilizing Large Language Models for DiseasePhenotyping in Obstructive Sleep Apnea

Learning From Aggregated Opinion

Rule Learning is Like Analogical Learning

Bilingualism is highly heterogeneous, in terms of language background, age of acquisition, and learning environment... Mapping these variations to brain structure is important to understand the relationship between language and brain. However, these differences in neuroanatomy are entangled with the variation from individual differences unrelated to bilingualism, for example, age and sex. Our goal is to disentangle bilingual-specific neuroanatomical variation from the common variation. First, we build a dataset including 12,000+ brain 3D images from bilinguals and monolinguals along with demographic and linguistic measures. Second, we train a contrastive neural network (Contrastive Variational Autoencoder, CVAE) on the dataset to learn the variations from bilingual-specific variations from variations shared with monolingual groups. We examine if the bilingual-specific neuroanatomical variations correlate with individual differences in linguistic measures through Representational Similarity Analysis. The initial results reveal that the common variation in brain structure shared by two groups can be identified by CVAE.

Contrastive neural network reveals the structure of neuroanatomical variation within bilingualism

Androids that strongly but imperfectly resemble humans in shape can elicit negative emotions in people, a phenomenon known as the "uncanny valley," which has been replicated in laboratory experiments. Recently, the accuracy of face recognition utilizing machine learning has increased, raising the question of whether machine learning can replicate the uncanny valley. Using FaceNet as a representative face recognition algorithm, we examined the similarity of face recognition to human evaluation and its replication of the uncanny valley. The results revealed a strong correlation between machine learning and human evaluation of human-like shapes. However, it is evident that only certain aspects of the uncanny valley were replicated. Furthermore, visualization of the activation maps suggests that localized regions, such as the mouth and chin, acted as the basis for judgment. These findings support the idea that human and machine learning have distinct areas of attention.

Does Machine Learning Replicate the Uncanny Valley? An Example using FaceNet

When we exercise sequences of actions, their execution becomes more fluent and precise. Here, we consider the possibility that exercised action sequences can also be used to make planning faster and more accurate by focusing expansion of the search tree on paths that have been frequently used in the past, and by reducing deep planning problems to shallow ones via multi-step jumps in the tree. To capture such sequences, we use a flexible Bayesian action chunking mechanism which finds and exploits statistically reliable structure at different scales. This gives rise to shorter or longer routines that can be embedded into a Monte-Carlo tree search planner. We show the benefits of this scheme using a physical construction task patterned after tangrams.


Habits of Mind: Reusing Action Sequences for Efficient Planning

Modeling human navigation and route selection in complex environments has used path planning in the past and more recently, dynamical models. The current study modeled the navigational movement trajectories of human herder agents tasked with corralling an autonomous target agent into a specified containment zone. Analyses of trajectories of participants completing a first-person herding game revealed that their movement entailed sequential (i) approach and then (ii) corral phases that were invariant to the initial location of the herder and target agents. More importantly, the trajectories of human herders could be captured using a low-dimensional environmentally coupled dynamical model. The novelty of this work is in extending the previous research in modeling navigational behaviors of humans to include moving targets that need to be manipulated into a desired position. Implications of the findings for the development of artificial agents for human-machine herding, as well as more general human-machine teaming are discussed.

Modeling Human Navigation in First-Person Herding Tasks

The ability to discriminate between large and small quantities is a core aspect of basic numerical competence in both humans and animals. In this work, we examine the extent to which the state-of-the-art neural networks designed for vision exhibit this basic ability. Motivated by studies in animal and infant numerical cognition, we use the numerical bisection procedure to test number discrimination in different families of neural architectures. Our results suggest that vision-specific inductive biases are helpful in numerosity discrimination, as models with such biases have lowest test errors on the task, and often have psychometric curves that qualitatively resemble those of humans and animals performing the task. However, even the strongest models, as measured on standard metrics of performance, fail to discriminate quantities in transfer experiments with differing training and testing conditions, indicating that such inductive biases might not be sufficient.

Evaluating Visual Number Discrimination in Deep Neural Networks

Recent work in computer vision and cognitive reasoning has given rise to an increasing adoption of the Violation-of-Expectation (VoE) paradigm in synthetic datasets. Inspired by infant psychology, researchers are now evaluating a model’s ability to label scenes as either expected or surprising with knowledge of only expected scenes. However, existing VoE-based 3D datasets in physical reasoning provide mainly vision data with little to no heuristics or inductive biases. Cognitive models of physical reasoning reveal infants create high-level abstract representations of objects and interactions. Capitalizing on this knowledge, we established a benchmark to study physical reasoning by curating a novel large-scale synthetic 3D VoE dataset armed with ground-truth heuristic labels of causally relevant features and rules. To validate our dataset in five event categories of physical reasoning, we benchmarked and analyzed human performance. We also proposed the Object File Physical Reasoning Network (OFPR-Net) which exploits the dataset's novel heuristics to outperform our baseline and ablation models. The OFPR-Net is also flexible in learning an alternate physical reality, showcasing its ability to learn universal causal relationships in physical reasoning to create systems with better interpretability.

A Benchmark for Modeling Violation-of-Expectation in Physical Reasoning Across Event Categories

We present a novel method for image saliency prediction that leverages a cognitive model of visual attention as an inductive bias.
This is in stark contrast to recent purely data-driven
models that have achieved performance improvements mainly by increased model capacity, resulting in high computational costs and the need for large scale, domain specific training data. 
We demonstrate that by leveraging a cognitive model of visual attention, our method achieves competitive performance to the state-of-the-art across several natural image datasets while only requiring a fraction of the parameters.
Furthermore, we set the new state of the art for saliency prediction on information visualizations, demonstrating the effectiveness of our approach for cross-domain generalization.
We further provide large-scale cognitively plausible synthetic gaze data on corresponding images in the full MSCOCO and FigureQA datasets, which we used for pre-training.
These results are highly promising and underline the significant potential of bridging between first-principle cognitive and data-driven models for computer vision tasks, potentially also beyond saliency prediction, and even visual attention.

\bf Keywords: neural networks, computational models of cognition, computer vision, saliency prediction, human visual attention

Improving neural saliency prediction with a cognitive model of human visual attention

Dependency tree is efficient for relation extraction model to exploit relations between words. Recent approaches have achieved promising performance while still suffering from inherent limitations. 
In this paper, we propose a novel relation extraction framework called Progressive Graph Learning over pruned dependency tree (PGLNet). PGLNet constructs a set of graphs by progressively adapting to input sentence. Specially, we implement the model to construct progressive weighted adjacency matrices by learning the relations among graph nodes with self-attention mechanism. Then, the model takes the learned weights as reference to prune dependency tree. Next, progressive convolution module is designed to encode the relations of entities and followed by relations classification. We evaluate our proposed model using public real-world datasets, experimental results demonstrate that the proposed model achieves state-of-the-art performance with consistency in all datasets. We conclude that the ability of PGLNet to progressively adapt to input data and enable the model with robustness.


Underline digital video library

Maximizing the Success Probability of Policy Allocations in Online Systems

Utilizing Large Language Models for DiseasePhenotyping in Obstructive Sleep Apnea

Learning From Aggregated Opinion

Rule Learning is Like Analogical Learning

Contrastive neural network reveals the structure of neuroanatomical variation within bilingualism

Does Machine Learning Replicate the Uncanny Valley? An Example using FaceNet

Habits of Mind: Reusing Action Sequences for Efficient Planning

Modeling Human Navigation in First-Person Herding Tasks

Evaluating Visual Number Discrimination in Deep Neural Networks

A Benchmark for Modeling Violation-of-Expectation in Physical Reasoning Across Event Categories

Improving neural saliency prediction with a cognitive model of human visual attention

Progressive Graph Learning over Pruned Dependency Trees For Relation Extraction

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES