Fangfang Xia

Matthias Eisenmann, Annika Reinke, Vivienn Weru et al. · utoronto

The number of international benchmarking competitions is steadily increasing in various fields of machine learning (ML) research and practice. So far, however, little is known about the common practice as well as bottlenecks faced by the community in tackling the research questions posed. To shed light on the status quo of algorithm development in the specific field of biomedical imaging analysis, we designed an international survey that was issued to all participants of challenges conducted in conjunction with the IEEE ISBI 2021 and MICCAI 2021 conferences (80 competitions in total). The survey covered participants' expertise and working environments, their chosen strategies, as well as algorithm characteristics. A median of 72% challenge participants took part in the survey. According to our results, knowledge exchange was the primary incentive (70%) for participation, while the reception of prize money played only a minor role (16%). While a median of 80 working hours was spent on method development, a large portion of participants stated that they did not have enough time for method development (32%). 25% perceived the infrastructure to be a bottleneck. Overall, 94% of all solutions were deep learning-based. Of these, 84% were based on standard architectures. 43% of the respondents reported that the data samples (e.g., images) were too large to be processed at once. This was most commonly addressed by patch-based training (69%), downsampling (37%), and solving 3D analysis tasks as a series of 2D tasks. K-fold cross-validation on the training set was performed by only 37% of the participants and only 50% of the participants performed ensembling based on multiple identical models (61%) or heterogeneous models (39%). 48% of the respondents applied postprocessing steps.

Xuefeng Liu, Fangfang Xia, Rick L. Stevens et al.

While training models and labeling data are resource-intensive, a wealth of pre-trained models and unlabeled data exists. To effectively utilize these resources, we present an approach to actively select pre-trained models while minimizing labeling costs. We frame this as an online contextual active model selection problem: At each round, the learner receives an unlabeled data point as a context. The objective is to adaptively select the best model to make a prediction while limiting label requests. To tackle this problem, we propose CAMS, a contextual active model selection algorithm that relies on two novel components: (1) a contextual model selection mechanism, which leverages context information to make informed decisions about which model is likely to perform best for a given context, and (2) an active query component, which strategically chooses when to request labels for data points, minimizing the overall labeling cost. We provide rigorous theoretical analysis for the regret and query complexity under both adversarial and stochastic settings. Furthermore, we demonstrate the effectiveness of our algorithm on a diverse collection of benchmark classification tasks. Notably, CAMS requires substantially less labeling effort (less than 10%) compared to existing methods on CIFAR10 and DRIFT benchmarks, while achieving similar or better accuracy. Our code is publicly available at: https://github.com/xuefeng-cs/Contextual-Active-Model-Selection.

Chinedu Innocent Nwoye, Deepak Alapatt, Tong Yu et al.

Context-aware decision support in the operating room can foster surgical safety and efficiency by leveraging real-time feedback from surgical workflow analysis. Most existing works recognize surgical activities at a coarse-grained level, such as phases, steps or events, leaving out fine-grained interaction details about the surgical activity; yet those are needed for more helpful AI assistance in the operating room. Recognizing surgical actions as triplets of <instrument, verb, target> combination delivers comprehensive details about the activities taking place in surgical videos. This paper presents CholecTriplet2021: an endoscopic vision challenge organized at MICCAI 2021 for the recognition of surgical action triplets in laparoscopic videos. The challenge granted private access to the large-scale CholecT50 dataset, which is annotated with action triplet information. In this paper, we present the challenge setup and assessment of the state-of-the-art deep learning methods proposed by the participants during the challenge. A total of 4 baseline methods from the challenge organizers and 19 new deep learning algorithms by competing teams are presented to recognize surgical action triplets directly from surgical videos, achieving mean average precision (mAP) ranging from 4.2% to 38.1%. This study also analyzes the significance of the results obtained by the presented approaches, performs a thorough methodological comparison between them, in-depth result analysis, and proposes a novel ensemble method for enhanced recognition. Our analysis shows that surgical workflow analysis is not yet solved, and also highlights interesting directions for future research on fine-grained surgical activity recognition which is of utmost importance for the development of AI in surgery.

Matthias Eisenmann, Annika Reinke, Vivienn Weru et al.

International benchmarking competitions have become fundamental for the comparative performance assessment of image analysis methods. However, little attention has been given to investigating what can be learnt from these competitions. Do they really generate scientific progress? What are common and successful participation strategies? What makes a solution superior to a competing method? To address this gap in the literature, we performed a multi-center study with all 80 competitions that were conducted in the scope of IEEE ISBI 2021 and MICCAI 2021. Statistical analyses performed based on comprehensive descriptions of the submitted algorithms linked to their rank as well as the underlying participation strategies revealed common characteristics of winning solutions. These typically include the use of multi-task learning (63%) and/or multi-stage pipelines (61%), and a focus on augmentation (100%), image preprocessing (97%), data curation (79%), and postprocessing (66%). The "typical" lead of a winning team is a computer scientist with a doctoral degree, five years of experience in biomedical image analysis, and four years of experience in deep learning. Two core general development strategies stood out for highly-ranked teams: the reflection of the metrics in the method design and the focus on analyzing and handling failure cases. According to the organizers, 43% of the winning algorithms exceeded the state of the art but only 11% completely solved the respective domain problem. The insights of our study could help researchers (1) improve algorithm development strategies when approaching new problems, and (2) focus on open research questions revealed by this work.

Xuefeng Liu, Songhao Jiang, Xiaotian Duan et al.

Protein-ligand binding is the process by which a small molecule (drug or inhibitor) attaches to a target protein. Binding affinity, which characterizes the strength of biomolecular interactions, is essential for tackling diverse challenges in life sciences, including therapeutic design, protein engineering, enzyme optimization, and elucidating biological mechanisms. Much work has been devoted to predicting binding affinity over the past decades. Here, we review recent significant works, with a focus on methods, evaluation strategies, and benchmark datasets. We note growing use of both traditional machine learning and deep learning models for predicting binding affinity, accompanied by an increasing amount of data on proteins and small drug-like molecules. With improved predictive performance and the FDA's phasing out of animal testing, AI-driven in silico models, such as AI virtual cells (AIVCs), are poised to advance binding affinity prediction; reciprocally, progress in building binding affinity predictors can refine AIVCs. Future efforts in binding affinity prediction and AI-driven in silico models can enhance the simulation of temporal dynamics, cell-type specificity, and multi-omics integration to support more accurate and personalized outcomes.

Zixuan Zhao, Nathan Wycoff, Neil Getty et al.

The field of neuromorphic computing is in a period of active exploration. While many tools have been developed to simulate neuronal dynamics or convert deep networks to spiking models, general software libraries for learning rules remain underexplored. This is partly due to the diverse, challenging nature of efforts to design new learning rules, which range from encoding methods to gradient approximations, from population approaches that mimic the Bayesian brain to constrained learning algorithms deployed on memristor crossbars. To address this gap, we present Neko, a modular, extensible library with a focus on aiding the design of new learning algorithms. We demonstrate the utility of Neko in three exemplar cases: online local learning, probabilistic learning, and analog on-device learning. Our results show that Neko can replicate the state-of-the-art algorithms and, in one case, lead to significant outperformance in accuracy and speed. Further, it offers tools including gradient comparison that can help develop new algorithmic variants. Neko is an open source Python library that supports PyTorch and TensorFlow backends.

Austin Clyde, Tom Brettin, Alexander Partin et al.

By combining various cancer cell line (CCL) drug screening panels, the size of the data has grown significantly to begin understanding how advances in deep learning can advance drug response predictions. In this paper we train >35,000 neural network models, sweeping over common featurization techniques. We found the RNA-seq to be highly redundant and informative even with subsets larger than 128 features. We found the inclusion of single nucleotide polymorphisms (SNPs) coded as count matrices improved model performance significantly, and no substantial difference in model performance with respect to molecular featurization between the common open source MOrdred descriptors and Dragon7 descriptors. Alongside this analysis, we outline data integration between CCL screening datasets and present evidence that new metrics and imbalanced data techniques, as well as advances in data standardization, need to be developed.

Hamed Hemati, Lorenzo Pellegrini, Xiaotian Duan et al.

Continual learning (CL) provides a framework for training models in ever-evolving environments. Although re-occurrence of previously seen objects or tasks is common in real-world problems, the concept of repetition in the data stream is not often considered in standard benchmarks for CL. Unlike with the rehearsal mechanism in buffer-based strategies, where sample repetition is controlled by the strategy, repetition in the data stream naturally stems from the environment. This report provides a summary of the CLVision challenge at CVPR 2023, which focused on the topic of repetition in class-incremental learning. The report initially outlines the challenge objective and then describes three solutions proposed by finalist teams that aim to effectively exploit the repetition in the stream to learn continually. The experimental results from the challenge highlight the effectiveness of ensemble-based solutions that employ multiple versions of similar modules, each trained on different but overlapping subsets of classes. This report underscores the transformative potential of taking a different perspective in CL by employing repetition in the data stream to foster innovative strategy design.

Robinson Umeike, Neil Getty, Fangfang Xia et al.

Large language models (LLMs) have demonstrated immense capabilities in understanding textual data and are increasingly being adopted to help researchers accelerate scientific discovery through knowledge extraction (information retrieval), knowledge distillation (summarizing key findings and methodologies into concise forms), and knowledge synthesis (aggregating information from multiple scientific sources to address complex queries, generate hypothesis and formulate experimental plans). However, scientific data often exists in both visual and textual modalities. Vision language models (VLMs) address this by incorporating a pretrained vision backbone for processing images and a cross-modal projector that adapts image tokens into the LLM dimensional space, thereby providing richer multimodal comprehension. Nevertheless, off-the-shelf VLMs show limited capabilities in handling domain-specific data and are prone to hallucinations. We developed intelligent assistants finetuned from LLaVA models to enhance multimodal understanding in low-dose radiation therapy (LDRT)-a benign approach used in the treatment of cancer-related illnesses. Using multilingual data from 42,673 articles, we devise complex reasoning and detailed description tasks for visual question answering (VQA) benchmarks. Our assistants, trained on 50,882 image-text pairs, demonstrate superior performance over base models as evaluated using LLM-as-a-judge approach, particularly in reducing hallucination and improving domain-specific comprehension.

Aneeq Zia, Max Berniker, Rogerio Garcia Nespolo et al.

Robotic assisted (RA) surgery promises to transform surgical intervention. Intuitive Surgical is committed to fostering these changes and the machine learning models and algorithms that will enable them. With these goals in mind we have invited the surgical data science community to participate in a yearly competition hosted through the Medical Imaging Computing and Computer Assisted Interventions (MICCAI) conference. With varying changes from year to year, we have challenged the community to solve difficult machine learning problems in the context of advanced RA applications. Here we document the results of these challenges, focusing on surgical tool localization (SurgToolLoc). The publicly released dataset that accompanies these challenges is detailed in a separate paper arXiv:2501.09209 [1].

Agastya P. Bhati, Shunzhou Wan, Dario Alfè et al.

The race to meet the challenges of the global pandemic has served as a reminder that the existing drug discovery process is expensive, inefficient and slow. There is a major bottleneck screening the vast number of potential small molecules to shortlist lead compounds for antiviral drug development. New opportunities to accelerate drug discovery lie at the interface between machine learning methods, in this case developed for linear accelerators, and physics-based methods. The two in silico methods, each have their own advantages and limitations which, interestingly, complement each other. Here, we present an innovative infrastructural development that combines both approaches to accelerate drug discovery. The scale of the potential resulting workflow is such that it is dependent on supercomputing to achieve extremely high throughput. We have demonstrated the viability of this workflow for the study of inhibitors for four COVID-19 target proteins and our ability to perform the required large-scale calculations to identify lead antiviral compounds through repurposing on a variety of supercomputers.

Alexander Partin, Thomas Brettin, Yvonne A. Evrard et al.

Motivated by the size of cell line drug sensitivity data, researchers have been developing machine learning (ML) models for predicting drug response to advance cancer treatment. As drug sensitivity studies continue generating data, a common question is whether the proposed predictors can further improve the generalization performance with more training data. We utilize empirical learning curves for evaluating and comparing the data scaling properties of two neural networks (NNs) and two gradient boosting decision tree (GBDT) models trained on four drug screening datasets. The learning curves are accurately fitted to a power law model, providing a framework for assessing the data scaling behavior of these predictors. The curves demonstrate that no single model dominates in terms of prediction performance across all datasets and training sizes, suggesting that the shape of these curves depends on the unique model-dataset pair. The multi-input NN (mNN), in which gene expressions and molecular drug descriptors are input into separate subnetworks, outperforms a single-input NN (sNN), where the cell and drug features are concatenated for the input layer. In contrast, a GBDT with hyperparameter tuning exhibits superior performance as compared with both NNs at the lower range of training sizes for two of the datasets, whereas the mNN performs better at the higher range of training sizes. Moreover, the trajectory of the curves suggests that increasing the sample size is expected to further improve prediction scores of both NNs. These observations demonstrate the benefit of using learning curves to evaluate predictors, providing a broader perspective on the overall data scaling characteristics. The fitted power law curves provide a forward-looking performance metric and can serve as a co-design tool to guide experimental biologists and computational scientists in the design of future experiments.

Yitan Zhu, Thomas Brettin, Yvonne A. Evrard et al.

Transfer learning has been shown to be effective in many applications in which training data for the target problem are limited but data for a related (source) problem are abundant. In this paper, we apply transfer learning to the prediction of anti-cancer drug response. Previous transfer learning studies for drug response prediction focused on building models that predict the response of tumor cells to a specific drug treatment. We target the more challenging task of building general prediction models that can make predictions for both new tumor cells and new drugs. We apply the classic transfer learning framework that trains a prediction model on the source dataset and refines it on the target dataset, and extends the framework through ensemble. The ensemble transfer learning pipeline is implemented using LightGBM and two deep neural network (DNN) models with different architectures. Uniquely, we investigate its power for three application settings including drug repurposing, precision oncology, and new drug development, through different data partition schemes in cross-validation. We test the proposed ensemble transfer learning on benchmark in vitro drug screening datasets, taking one dataset as the source domain and another dataset as the target domain. The analysis results demonstrate the benefit of applying ensemble transfer learning for predicting anti-cancer drug response in all three applications with both LightGBM and DNN models. Compared between the different prediction models, a DNN model with two subnetworks for the inputs of tumor features and drug features separately outperforms LightGBM and the other DNN model that concatenates tumor features and drug features for input in the drug repurposing and precision oncology applications. In the more challenging application of new drug development, LightGBM performs better than the other two DNN models.

Neil Getty, Zixuan Zhao, Stephan Gruessner et al.

Robot-assisted minimally invasive surgery is improving surgeon performance and patient outcomes. This innovation is also turning what has been a subjective practice into motion sequences that can be precisely measured. A growing number of studies have used machine learning to analyze video and kinematic data captured from surgical robots. In these studies, models are typically trained on benchmark datasets for representative surgical tasks to assess surgeon skill levels. While they have shown that novices and experts can be accurately classified, it is not clear whether machine learning can separate highly proficient surgeons from one another, especially without video data. In this study, we explore the possibility of using only kinematic data to predict surgeons of similar skill levels. We focus on a new dataset created from surgical exercises on a simulation device for skill training. A simple, efficient encoding scheme was devised to encode kinematic sequences so that they were amenable to edge learning. We report that it is possible to identify surgical fellows receiving near perfect scores in the simulation exercises based on their motion characteristics alone. Further, our model could be converted to a spiking neural network to train and infer on the Nengo simulation framework with no loss in accuracy. Overall, this study suggests that building neuromorphic models from sparse motion features may be a potentially useful strategy for identifying surgeons and gestures with chips deployed on robotic systems to offer adaptive assistance during surgery and training with additional latency and privacy benefits.

Neil Getty, Thomas Brettin, Dong Jin et al.

We explore three representative lines of research and demonstrate the utility of our methods on a classification benchmark of brain cancer MRI data. First, we present a capsule network that explicitly learns a representation robust to rotation and affine transformation. This model requires less training data and outperforms both the original convolutional baseline and a previous capsule network implementation. Second, we leverage the latest domain adaptation techniques to achieve a new state-of-the-art accuracy. Our experiments show that non-medical images can be used to improve model performance. Finally, we design a spiking neural network trained on the Intel Loihi neuromorphic chip (Fig. 1 shows an inference snapshot). This model consumes much lower power while achieving reasonable accuracy given model reduction. We posit that more research in this direction combining hardware and learning advancements will power future medical imaging (on-device AI, few-shot prediction, adaptive scanning).

Nathan Wycoff, Prasanna Balaprakash, Fangfang Xia

If edge devices are to be deployed to critical applications where their decisions could have serious financial, political, or public-health consequences, they will need a way to signal when they are not sure how to react to their environment. For instance, a lost delivery drone could make its way back to a distribution center or contact the client if it is confused about how exactly to make its delivery, rather than taking the action which is "most likely" correct. This issue is compounded for health care or military applications. However, the brain-realistic temporal credit assignment problem neuromorphic computing algorithms have to solve is difficult. The double role weights play in backpropagation-based-learning, dictating how the network reacts to both input and feedback, needs to be decoupled. e-prop 1 is a promising learning algorithm that tackles this with Broadcast Alignment (a technique where network weights are replaced with random weights during feedback) and accumulated local information. We investigate under what conditions the Bayesian loss term can be expressed in a similar fashion, proposing an algorithm that can be computed with only local information as well and which is thus no more difficult to implement on hardware. This algorithm is exhibited on a store-recall problem, which suggests that it can learn good uncertainty on decisions to be made over time.

Prasanna Balaprakash, Romain Egele, Misha Salim et al.

Cancer is a complex disease, the understanding and treatment of which are being aided through increases in the volume of collected data and in the scale of deployed computing power. Consequently, there is a growing need for the development of data-driven and, in particular, deep learning methods for various tasks such as cancer diagnosis, detection, prognosis, and prediction. Despite recent successes, however, designing high-performing deep learning models for nonimage and nontext cancer data is a time-consuming, trial-and-error, manual task that requires both cancer domain and deep learning expertise. To that end, we develop a reinforcement-learning-based neural architecture search to automate deep-learning-based predictive model development for a class of representative cancer data. We develop custom building blocks that allow domain experts to incorporate the cancer-data-specific characteristics. We show that our approach discovers deep neural network architectures that have significantly fewer trainable parameters, shorter training time, and accuracy similar to or higher than those of manually designed architectures. We study and demonstrate the scalability of our approach on up to 1,024 Intel Knights Landing nodes of the Theta supercomputer at the Argonne Leadership Computing Facility.

Nathan Wycoff, Prasanna Balaprakash, Fangfang Xia

As neural networks have begun performing increasingly critical tasks for society, ranging from driving cars to identifying candidates for drug development, the value of their ability to perform uncertainty quantification (UQ) in their predictions has risen commensurately. Permanent dropout, a popular method for neural network UQ, involves injecting stochasticity into the inference phase of the model and creating many predictions for each of the test data. This shifts the computational and energy burden of deep neural networks from the training phase to the inference phase. Recent work has demonstrated near-lossless conversion of classical deep neural networks to their spiking counterparts. We use these results to demonstrate the feasibility of conducting the inference phase with permanent dropout on spiking neural networks, mitigating the technique's computational and energy burden, which is essential for its use at scale or on edge platforms. We demonstrate the proposed approach via the Nengo spiking neural simulator on a combination drug therapy dataset for cancer treatment, where UQ is critical. Our results indicate that the spiking approximation gives a predictive distribution practically indistinguishable from that given by the classical network.

John W. Santerre, James J. Davis, Fangfang Xia et al.

Biological datasets amenable to applied machine learning are more available today than ever before, yet they lack adequate representation in the Data-for-Good community. Here we present a work in progress case study performing analysis on antimicrobial resistance (AMR) using standard ensemble machine learning techniques and note the successes and pitfalls such work entails. Broadly, applied machine learning (AML) techniques are well suited to AMR, with classification accuracies ranging from mid-90% to low- 80% depending on sample size. Additionally, these techniques prove successful at identifying gene regions known to be associated with the AMR phenotype. We believe that the extensive amount of biological data available, the plethora of problems presented, and the global impact of such work merits the consideration of the Data- for-Good community.