In this paper, we take an inspiration from human perception and explore a compositional approach for egocentric video representation.

Neural semantic parsers have shown good overall performance for a variety of linguistic phenomena, reaching semantic matching scores of more than 90%. But how do such parsers perform on strongly context-sensitive phenomena, where large pieces of semantic information need to be duplicated to form a meaningful semantic representation? A case in point is English verb phrase ellipsis, a construct where entire verb phrases can be abbreviated by a single auxiliary verb. Are the otherwise known as powerful semantic parsers able to deal with ellipsis or aren't they? We constructed a corpus of 120 cases of ellipsis with their fully resolved meaning representation and used this as a challenge set for a large battery of neural semantic parsers. Although these parsers performed very well on the standard test set, they failed in the instances with ellipsis. Data augmentation helped improve the parsing results. The reason for the difficulty of parsing elided phrases is not that copying semantic material is hard, but that usually occur in linguistically complicated contexts causing most of the parsing errors.

Text-video retrieval (TVR) systems often suffer from visual-linguistic biases present in datasets, which cause pre-trained vision-language models to overlook key details. To address this, we propose BiMa, a novel framework designed to mitigate biases in both visual and textual representations. Our approach begins by generating scene elements that characterize each video by identifying relevant entities/objects and activities. For visual debiasing, we integrate these scene elements into the video embeddings, enhancing them to emphasize fine-grained and salient details. For textual debiasing, we introduce a mechanism to disentangle text features into content and bias components, enabling the model to focus on meaningful content while separately handling biased information. Extensive experiments and ablation studies across five major TVR benchmarks (i.e., MSR-VTT, MSVD, LSMDC, ActivityNet, and DiDeMo) demonstrate the competitive performance of BiMa. Additionally, the model's bias mitigation capability is consistently validated by its strong results on out-of-distribution retrieval tasks.

Online social networks and micro-blogging services are no longer limited to the followers of the latest technologies or teenagers, as might once have been expected. Such technology and services are becoming widely adopted by the mainstream population as an integral part of their daily lives (Fox et al., 2009). A very prominent example of such an application is Twitter, a micro-blogging service. Twitter lets its users post very short (at most 140-character) messages - which are called tweets - about what they have been doing or thinking, or what they want to share with their friends and other people. Everyday, tens of millions of tweets are posted by users worldwide. The proliferation of publicly available, user-generated content is a vast source of social data and is already shaping the field of computational social science (Lazer et al., 2009; Thelwall et al., 2010a). Another field which enjoys the abundance of Web-based text is knowledge extraction and automated ontology building. An example application is KNEXT ( Kn owledge Ex traction from T ext) - a system proposed for extracting "general world knowledge from miscellaneous texts, including fiction" (Schubert and Tong, 2003). Web-based text is increasingly used as a source for everyday knowledge (frequently referred as commonsense knowledge).

Guided by grammatical structure, words compose to form sentences, and guided by discourse structure, sentences compose to form dialogues and documents. The compositional aspect of sentence and discourse units is often overlooked by machine learning algorithms. A recent initiative called Quantum Natural Language Processing (QNLP) learns word meanings as points in a Hilbert space and acts on them via a translation of grammatical structure into Parametrised Quantum Circuits (PQCs). Previous work extended the QNLP translation to discourse structure using points in a closure of Hilbert spaces. In this paper, we evaluate this translation on a Winograd-style pronoun resolution task. We train a Variational Quantum Classifier (VQC) for binary classification and implement an end-to-end pronoun resolution system. The simulations executed on IBMQ software converged with an F1 score of 87.20%. The model outperformed two out of three classical coreference resolution systems and neared state-of-the-art SpanBERT. A mixed quantum-classical model yet improved these results with an F1 score increase of around 6%.

Detecting contradictions in text is one of the hardest tasks The intuition is to use linguistic and factual rules where for a language model to comprehend. This is due to the this is applicable, i.e. for contradictions based on antonymy, complex semantic nature of contradictions, and the variety of negations and numeric mismatches. For more complex relations contexts in which they can occur. For this reason, a multitude such as factive or structural contradictions, we instruct of data sets and models have been developed to solve this a generative model to produce new samples, either based task. Meanwhile, the recent onset of large generative language on given premises or on the type description alone. To our models has given rise to new possibilities for problem solving knowledge, this is the first work implementing such a hybrid as well as data augmentation, which we aim to explore in this data generation method with respect to NLI.

Automaton-based representations of task knowledge play an important role in control and planning for sequential decision-making problems. However, obtaining the high-level task knowledge required to build such automata is often difficult. Meanwhile, large-scale generative language models (GLMs) can automatically generate relevant task knowledge. However, the textual outputs from GLMs cannot be formally verified or used for sequential decision-making. We propose a novel algorithm named GLM2FSA, which constructs a finite state automaton (FSA) encoding high-level task knowledge from a brief natural-language description of the task goal. GLM2FSA first sends queries to a GLM to extract task knowledge in textual form, and then it builds an FSA to represent this text-based knowledge. The proposed algorithm thus fills the gap between natural-language task descriptions and automaton-based representations, and the constructed FSA can be formally verified against user-defined specifications. We accordingly propose a method to iteratively refine the queries to the GLM based on the outcomes, e.g., counter-examples, from verification. We demonstrate GLM2FSA's ability to build and refine automaton-based representations of everyday tasks (e.g., crossing a road), and also of tasks that require highly-specialized knowledge (e.g., executing secure multi-party computation).

Recently developed pretrained models can encode rich world knowledge expressed in multiple modalities, such as text and images. However, the outputs of these models cannot be integrated into algorithms to solve sequential decision-making tasks. We develop an algorithm that utilizes the knowledge from pretrained models to construct and verify controllers for sequential decision-making tasks, and to ground these controllers to task environments through visual observations. In particular, the algorithm queries a pretrained model with a user-provided, text-based task description and uses the model's output to construct an automaton-based controller that encodes the model's task-relevant knowledge. It then verifies whether the knowledge encoded in the controller is consistent with other independently available knowledge, which may include abstract information on the environment or user-provided specifications. If this verification step discovers any inconsistency, the algorithm automatically refines the controller to resolve the inconsistency. Next, the algorithm leverages the vision and language capabilities of pretrained models to ground the controller to the task environment. It collects image-based observations from the task environment and uses the pretrained model to link these observations to the text-based control logic encoded in the controller (e.g., actions and conditions that trigger the actions). We propose a mechanism to ensure the controller satisfies the user-provided specification even when perceptual uncertainties are present. We demonstrate the algorithm's ability to construct, verify, and ground automaton-based controllers through a suite of real-world tasks, including daily life and robot manipulation tasks.

Despite the population of speakers, Yorùbá is still considered as a low The advancement in Natural language resource language (for which few language Processing (NLP) can be attributed to recent resources exist), making it very difficult for the improvements in the strategy and techniques of development of more advanced models such as the large data collection, archiving, analysis, and Neural Machine model that requires large volumes visualization. NLP began in the '50s as machine of data. With the number of speakers, translating translation (MT), intended to aid in code-breaking the language to other widely spoken languages was during World War II although the translations were not initially emphasized. However, recent not successful, these early stages of MT were linguistic researchers are taking up the challenges necessary stepping stones on the way to more by giving more attention (as compared to the highresource sophisticated technologies (Zhang, 2018; Quinn, language of the Western World).