AI Technology & Industry Review
In the new paper Fine-tuning Language Models To Find Agreement Among Humans With Diverse Preferences, a research team from DeepMind and University College London fine-tunes a 70 billion parameter language model to generate statements that maximize agreement among a human group with diverse written opinions.
In the new paper Solving Math Word Problems With Process- and Outcome-based Feedback, a DeepMind research team conducts the first comprehensive comparison between process- and outcome-based model supervision. The two approaches achieve comparable final-answer error rate improvements on math word problems, while the process-based method significantly reduces reasoning errors from 14.0 to just 3.4 percent.
In the new paper Fixing Model Bugs with Natural Language Patches, researchers from Stanford University and Microsoft Research propose a method that uses declarative statements as feedback for correcting errors in neural models, significantly increasing accuracy without high compute costs.
In the new paper Fast DistilBERT on CPUs, researchers from Intel Corporation and Intel Labs propose a pipeline and hardware-aware extreme compression technique for creating and running fast transformer models on CPUs. The approach achieves impressive speed ups and SOTA performance in production environments.
In the new paper Fine-Tuning Language Models via Epistemic Neural Networks, a DeepMind research team modifies large language models to create an Epistemic Neural Network. The novel approach achieves model performance comparable to that obtained via fine-tuning while requiring 50 percent less data.
In the new paper Locating and Editing Factual Associations in GPT, a research team from MIT CSAIL, Northeastern University and Technion IIT examines how information flows during knowledge recall in large autoregressive transformers and introduces Rank-One Model Editing (ROME), a simple, zero-shot principled model editor capable of locating and editing factual associations in such models.
In the new paper Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Them, a Google Research and Stanford University team applies chain-of-thought (CoT) prompting — a series of intermediate reasoning steps — to 23 BIG-Bench tasks on which language models have failed to outperform the average human rater. The proposed approach enables models to surpass human performance on 17 of the 23 tasks.
In the new paper Ask Me Anything: A Simple Strategy for Prompting Language Models, a research team from Stanford University, Numbers Station, and the University of Wisconsin-Madison presents Ask Me Anything Prompting (AMA), a simple large language model prompting strategy that enables a 30x smaller language model to outperform few-shot GPT3-175B.
In the new paper Promptagator: Few-shot Dense Retrieval From 8 Examples, a Google Research team proposes Prompt-based Query Generation for Retriever (Promptagator), a novel and simple approach for few-shot retrieval that leverages large language model (LLM) prompting to generate synthetic task-specific training data.
In the new paper Vec2text With Round-Trip Translations, Google Brain researchers explore large language models’ capabilities for generating arbitrary natural language text from inputs of fixed-size vectors — a vec2text setting — and propose a simple data augmentation approach based on round-trip translations to improve vec2text model performance.
Colossal-AI, a unified deep learning system for the big model era, can efficiently and rapidly deploy large AI model training and inference with just a few lines of code, and promote the low-cost application and implementation of big models.
In the new paper Faithful Reasoning Using Large Language Models, a DeepMind research team proposes a forward-chaining selection-inference model that performs faithful reasoning and provides a valid reasoning trace to improve reasoning quality and help users validate the model’s final answers.
In the new paper PEER: A Collaborative Language Model, a research team from Meta AI, Carnegie Mellon University, PSL University, and University College London presents PEER, a collaborative language model that performs a humanlike writing process — composing drafts, adding suggestions, proposing edits and providing explanations for its actions.
In the new paper Z-Code++: A Pre-trained Language Model Optimized for Abstractive Summarization, a research team from Microsoft Azure AI and Microsoft Research presents Z-Code++, a novel encoder-decoder pretrained language model optimized for abstractive summarization that significantly improves performance on low-resource summarization tasks.
In the new paper Efficient Training of Language Models to Fill in the Middle, an OpenAI research team shows that causal decoder-based autoregressive (AR) language models can learn to infill texts via a very simple and straightforward transformation to the training data and without any architectural modifications.
In the new paper Confident Adaptive Language Modeling, a research team from Google and MIT presents Confident Adaptive Language Modeling (CALM), a framework that dynamically allocates different amounts of compute to each input and generation timestep, achieving up to 3x speedups while maintaining high performance.
In the new paper Evolution Through Large Models, an OpenAI research team shows that large-scale language models (LLMs) trained to generate modern programming language can suggest intelligent mutations that can be leveraged to realize dramatically improved mutation operators for genetic programming.
In the new paper LegoNN: Building Modular Encoder-Decoder Models, Meta AI researchers propose LegoNN, a procedure for building encoder-decoder architectures with decoder modules that can be shared across different tasks without finetuning or significant performance reductions.
In the new paper Beyond the Imitation Game: Quantifying and Extrapolating the Capabilities of Language Models, 444 authors from 132 institutions introduce Beyond the Imitation Game (BIG-bench), a large-scale, extremely difficult and diverse benchmark that includes 204 tasks for predicting the potentially transformative effects of large language models.
In the new paper UViM: A Unified Modeling Approach for Vision with Learned Guiding Codes, a Google Brain research team proposes UViM, a unified approach that leverages language modelling and discrete representation learning to enable the modelling of a wide range of computer vision tasks without task-specific modifications.
In the new paper Tracing Knowledge in Language Models Back to the Training Data, a team from MIT CSAIL and Google Research proposes a benchmark for tracing language models’ assertions to the associated training data, aiming to establish a principled ground truth and mitigate high compute demands for large neural language model training.
In the new paper Large Language Models are Zero-Shot Reasoners, a research team from the University of Tokyo and Google Brain demonstrates that large language models (LLMs) can become good zero-shot reasoners through the addition of a simple prompt — “Let’s think step by step” — that elicits a step-by-step thinking process before each question is answered. Their Zero-shot-CoT model achieves huge performance gains compared to the zero-shot baseline.
In the new paper Standing on the Shoulders of Giant Frozen Language Models, AI21 Labs researchers propose three novel methods for learning small neural modules that specialize a frozen language model to different tasks. Their compute-saving approach outperforms conventional frozen model methods and challenges fine-tuning performance without sacrificing model versatility.
In the new paper Unifying Language Learning Paradigms, a Google Research/Brain team proposes a framework for pretraining universal language models that are effective across many different tasks. Their 20B parameter model surpasses 175B GPT-3 on the zero-shot SuperGLUE benchmark and triples the performance of T5-XXL on one-shot summarization tasks.
In the new paper Socratic Models: Composing Zero-Shot Multimodal Reasoning with Language, Google researchers argue that the diversity of different foundation models is symbiotic and that it is possible to build a framework that uses structured Socratic dialogue between pre-existing foundation models to formulate new multimodal tasks as a guided exchange between the models without additional finetuning.
A Stanford research team proposes Time Control (TC), a language model that implicitly plans via a latent stochastic process and generates texts consistent with this latent plan to improve performance on long text generation.
DeepMind trains agents to use keyboard and mouse commands with pixel and Document Object Model (DOM) observations to control computers, achieving state-of-the-art and human-level mean performance across all tasks on the MiniWob++ benchmark.
A research team from Alibaba Group’s DAMO Academy proposes OFA (One For All), a pretrained model that unifies modalities and tasks to a simple Seq2Seq learning framework and achieves SOTA results on a series of multimodal tasks.
An OpenAI research team presents an expert iteration-based neural theorem prover capable of solving a curriculum of increasingly difficult mathematical problems (such as high-school olympiad-level problems) from a set of formal statements of sufficiently varied difficulty and without the need for associated ground-truth proofs.
A research team from Microsoft and NVIDIA leverages the NVIDIA Megatron-LM and Microsoft’s DeepSpeed to create an efficient and scalable 3D parallel system that combines data, pipeline, and tensor-slicing based parallelism, achieving superior zero-, one-, and few-shot learning accuracies and new state-of-the-art results on NLP benchmarks.
An OpenAI research team leverages reinforcement learning from human feedback (RLHF) to make significant progress on aligning language models with the users’ intentions. The proposed InstructGPT models are better at following instructions than GPT-3 while also more truthful and less toxic.
A team from Google Research, University of Pennsylvania and Cornell University proposes a principled perspective to filter out common memorization for LMs, introducing “counterfactual memorization” to measure the expected change in a model’s prediction and distinguish “rare” (episodic) memorization from “common” (semantic) memorization in neural LMs.
In the new paper Towards Interactive Language Modeling, a research team from the University of Amsterdam and Meta AI Labs presents a road map detailing the steps to be taken towards interactive language modelling.
A DeepMind research team proposes RETRO (Retrieval-Enhanced Transformer), an enhanced auto-regressive language model that conditions on document chunks retrieved from a large corpus and achieves performance comparable to GPT-3 and Jurassic-1 on the Pile dataset while using 25× fewer parameters.
In the new paper Shaking the Foundations: Delusions in Sequence Models for Interaction and Control, a DeepMind research team explores the origins of mismatch problems in sequence models that lack understanding of the cause and effect of their actions, and addresses the problem by treating actions as causal interventions.
A Google AI research team explores zero-label learning (training with synthetic data only) in natural language processing, and introduces Unsupervised Data Generation (UDG), a training data creation procedure designed to synthesize high-quality training data without human annotations.
A new MIT CSAIL paper presents an alternative, hierarchical approach to sequence-to-sequence learning with quasi-synchronous grammars to improve sequence-to-sequence models’ compositional generalization on diagnostic tasks.
Researchers from Instituto de Telecomunicações, DeepMind, Institute of Systems and Robotics, Instituto Superior Técnico and Unbabel propose “∞-former” — a transformer model with unbounded long-term memory (LTM) that can attend to arbitrarily long contexts.
A Huawei research team proposes a topic-based personalized ranking model (TPRM) that integrates pretrained contextualized term representations and user profiles constructed by a topic model to tailor the output ranking list.
A Google Research team proposes Wordcraft, a text editor with a built-in AI-powered creative writing assistant. Wordcraft uses few-shot learning and the natural affordances of conversation to support a variety of user interactions; and can help with story planning, writing and editing.