Few-shot in-context learning (ICL) enables pre-trained language models to perform a previously-unseen task without any gradient-based training by feeding a small number of training examples as part of the input. ICL incurs substantial computational, memory, and storage costs because it involves processing all of the training examples every time a prediction is made. Parameter-efficient fine-tuning (PEFT) (e.g. adapter modules, prompt tuning, sparse update methods, etc.) offers an alternative paradigm where a small set of parameters are trained to enable a model to perform the new task. In this paper, we rigorously compare few-shot ICL and PEFT and demonstrate that the latter offers better accuracy as well as dramatically lower computational costs. Along the way, we introduce a new PEFT method called (IA)$^3$ that scales activations by learned vectors, attaining stronger performance while only introducing a relatively tiny amount of new parameters. We also propose a simple recipe based on the T0 model called T-Few that can be applied to new tasks without task-specific tuning or modifications. We validate the effectiveness of T-Few on completely unseen tasks by applying it to the RAFT benchmark, attaining super-human performance for the first time and outperforming the state-of-the-art by 6% absolute. All of the code used in our experiments is publicly available.

Partial-label learning is a kind of weakly-supervised learning with inexact labels, where for each training example, we are given a set of candidate labels instead of only one true label. Recently, various approaches on partial-label learning have been proposed under different generation models of candidate label sets. However, these methods require relatively strong distributional assumptions on the generation models. When the assumptions do not hold, the performance of the methods is not guaranteed theoretically. In this paper, we propose the notion of properness on partial labels. We show that this proper partial-label learning framework requires a weaker distributional assumption and includes many previous partial-label learning settings as special cases. We then derive a unified unbiased estimator of the classification risk. We prove that our estimator is risk-consistent, and we also establish an estimation error bound. Finally, we validate the effectiveness of our algorithm through experiments.

Tracking all nuclei of an embryo in noisy and dense fluorescence microscopy data is a challenging task. We build upon a recent method for nuclei tracking that combines weakly-supervised learning from a small set of nuclei center point annotations with an integer linear program (ILP) for optimal cell lineage extraction. Our work specifically addresses the following challenging properties of C. elegans embryo recordings: (1) Many cell divisions as compared to benchmark recordings of other organisms, and (2) the presence of polar bodies that are easily mistaken as cell nuclei. To cope with (1), we devise and incorporate a learnt cell division detector. To cope with (2), we employ a learnt polar body detector. We further propose automated ILP weights tuning via a structured SVM, alleviating the need for tedious manual set-up of a respective grid search. Our method outperforms the previous leader of the cell tracking challenge on the Fluo-N3DH-CE embryo dataset. We report a further extensive quantitative evaluation on two more C. elegans datasets. We will make these datasets public to serve as an extended benchmark for future method development. Our results suggest considerable improvements yielded by our method, especially in terms of the correctness of division event detection and the number and length of fully correct track segments. Code: https://github.com/funkelab/linajea

The majority of knowledge graph embedding techniques treat entities and predicates as separate embedding matrices, using aggregation functions to build a representation of the input triple. However, these aggregations are lossy, i.e. they do not capture the semantics of the original triples, such as information contained in the predicates. To combat these shortcomings, current methods learn triple embeddings from scratch without utilizing entity and predicate embeddings from pre-trained models. In this paper, we design a novel fine-tuning approach for learning triple embeddings by creating weak supervision signals from pre-trained knowledge graph embeddings. We develop a method for automatically sampling triples from a knowledge graph and estimating their pairwise similarities from pre-trained embedding models. These pairwise similarity scores are then fed to a Siamese-like neural architecture to fine-tune triple representations. We evaluate the proposed method on two widely studied knowledge graphs and show consistent improvement over other state-of-the-art triple embedding methods on triple classification and triple clustering tasks.

This article was originally an episode of MLOps Live, an interactive Q&A session where ML practitioners answer questions from other ML practitioners. Every episode is focused on one specific ML topic, and during this one, we talked to Mateusz Opala about leveraging unlabeled image data with self-supervised learning or pseudo-labeling. But, if you prefer a written version, here it is! Sabine: With us today, we have Mateusz Opala, who is going to be answering questions about leveraging unlabeled image data with self-supervised learning or pseudo-labeling. Sabine: It's great to have you. Mateusz has held a number of leading machine learning positions at companies like Netguru and Brainly. So, Mateusz, you have a background in computer science, but how did you get more into the machine learning side of things? Mateusz: It all started during my sophomore year at university. One of my professors told me that Andrew Ng was doing his first iteration of the famous course on machine learning on Coursera. I kind of started from there, then did a bachelor thesis on deep unsupervised learning and went to Siemens to work in deep learning, and then all my positions were strictly about machine learning. Sabine: You've been on that path ever since? I worked for some time before as a backend engineer. But for most of the time in my career, I was a machine learning engineer/data scientist. Sabine: Mateusz, to warm you up.

Recently, e-learning platforms have grown as a place where students can post doubts (as a snap taken with smart phones) and get them resolved in minutes. However, the significant increase in the number of student-posted doubts with high variance in quality on these platforms not only presents challenges for teachers' navigation to address them but also increases the resolution time per doubt. Both are not acceptable, as high doubt resolution time hinders the students learning progress. This necessitates ways to automatically identify if there exists a similar doubt in repository and then serve it to the teacher as the plausible solution to validate and communicate with the student. Supervised learning techniques (like Siamese architecture) require labels to identify the matches, which is not feasible as labels are scarce and expensive. In this work, we, thus, developed a label-agnostic doubt matching paradigm based on the representations learnt via self-supervised technique. Building on prior theoretical insights of BYOL (bootstrap your own latent space), we propose custom BYOL which combines domain-specific augmentation with contrastive objective over a varied set of appropriately constructed data views. Results highlighted that, custom BYOL improves the top-1 matching accuracy by approximately 6\% and 5\% as compared to both BYOL and supervised learning instances, respectively. We further show that both BYOL-based learning instances performs either on par or better than human labeling.

InstDisc 首次提出個體判別的代理任務(pretext task)，如果說 MOCO 是站在巨人的肩膀上 的話，InstDisc 就是那巨人的肩膀，InstDisc 就是 MOCO 中反覆提到的文獻[61]。除了 MOCO 以外，還有很多其他的論文和實驗的細節都是直接根據 InstDisc 而來的。 圖 1 展示了 InstDisc…

Attackers use malicious software, known as malware, to steal sensitive data, damage network infrastructure, and hold information for ransom. One of the top priorities for computer security tools is to detect malware and prevent or minimize its impact on both corporate and personal networks. Traditionally, signature-based methods have been used to detect files previously identified as malicious with near perfect precision, but potentially miss newer malware samples. With the advent of self-modifying malware and the rapid increase in novel threats, signature-based methods are insufficient on their own. By generalizing patterns of known benign/malicious training examples, machine learning (ML) exhibits the capability to quickly and accurately classify novel file samples in many research studies [19]. Moreover, ML-based malware research has made the transition from the subject of myriad research efforts to a current mainstay of commercial-off-the-shelf (COTS) malware detectors. Yet, few practical evaluations of COTS ML-based technologies have been conducted. Turning from the academic literature to market reports from commercial companies can provide (for a fee) useful information, specifically, end-user feedback, itemization of all technologies in the antivirus/endpoint detection and response marketplace [17], and even statistics showing the efficacy of the detectors on malware tests [4, 40].

Weak supervision has been applied to various Natural Language Understanding tasks in recent years. Due to technical challenges with scaling weak supervision to work on long-form documents, spanning up to hundreds of pages, applications in the document understanding space have been limited. At Lexion, we built a weak supervision-based system tailored for long-form (10-200 pages long) PDF documents. We use this platform for building dozens of language understanding models and have applied it successfully to various domains, from commercial agreements to corporate formation documents. In this paper, we demonstrate the effectiveness of supervised learning with weak supervision in a situation with limited time, workforce, and training data. We built 8 high quality machine learning models in the span of one week, with the help of a small team of just 3 annotators working with a dataset of under 300 documents. We share some details about our overall architecture, how we utilize weak supervision, and what results we are able to achieve. We also include the dataset for researchers who would like to experiment with alternate approaches or refine ours. Furthermore, we shed some light on the additional complexities that arise when working with poorly scanned long-form documents in PDF format, and some of the techniques that help us achieve state-of-the-art performance on such data.

Frozen models trained to mimic static datasets can never improve their performance. Models that can employ internet-retrieval for up-to-date information and obtain feedback from humans during deployment provide the promise of both adapting to new information, and improving their performance. In this work we study how to improve internet-driven conversational skills in such a learning framework. We collect deployment data, which we make publicly available, of human interactions, and collect various types of human feedback -- including binary quality measurements, free-form text feedback, and fine-grained reasons for failure. We then study various algorithms for improving from such feedback, including standard supervised learning, rejection sampling, model-guiding and reward-based learning, in order to make recommendations on which type of feedback and algorithms work best. We find the recently introduced Director model (Arora et al., '22) shows significant improvements over other existing approaches.