Northeastern University London Repository
Not a member yet
1286 research outputs found
Sort by
A Simplified Fish School Search Algorithm for Continuous Single Objective Optimisation
Full text linkThe Fish School Search (FSS) algorithm is a metaheuristic known for its distinctive exploration and exploitation operators and cumulative success representation approach. Despite its success across various problem domains, FSS presents issues due to its high number of parameters, making its performance susceptible to improper parameterisation. Additionally, the interplay between its operators requires a sequential execution in a specific order, requiring two fitness evaluations per iteration for each individual. This operator's intricacy and the number of fitness evaluations pose the issue of costly fitness functions and inhibit parallelisation. To address these challenges, this paper proposes a Simplified Fish School Search (SFSS) algorithm that preserves the core features of the original FSS while redesigning the fish movement operators and introducing a new turbulence mechanism to enhance population diversity and robustness against stagnation. The SFSS also reduces the number of fitness evaluations per iteration and minimises the algorithm's parameter set. Computational experiments were conducted using a benchmark suite from the CEC 2017 competition to compare the SFSS with the traditional FSS and five other well-known metaheuristics. The SFSS outperformed the FSS in 84\% of the problems, and achieved the best results among all algorithms in 10 of the 26 problems
Countering Nationalist Politics Through Distant Witnessing: An Affective-Discursive Approach
Full text linkAdvancing an affective-discursive approach, this article examines how distant witnessing bolsters digital activist organizing that challenges a Southern authoritarian regime’s nationalist politics on international social media. The research foregrounds an initiative coordinated by a group of overseas dissidents who sought to expose China’s foreign policy regarding the Russian invasion of Ukraine on X (Twitter). Amid ongoing battlefield stalemates that have prolonged the war, this initiative has evolved into a broader form of resistance, staging distant witnessing through X postings to circumvent the party-state’s internet censorship. As this process unfolds, activists capitalize on social media affordances to curate witness testimonies that counter China’s distorted geopolitical narratives and beyond. Resorting to both pre-discursive and discursive gestures, this repertoire contributes to the formation of affective publics, thereby stimulating public contention over the party-state’s authoritarian rule from afar. The findings highlight how such digital platforms as X enable an informal model of organization, allowing Southern activists to act as mediators who coordinate distant witnessing in a transnational context
Mapping the Proceedings: The Importance of Spatiality for Reconstructing Black and Multiracial Communities in Georgian London
No full textExploring the vast references to spatiality in eighteenth-century criminal justice records reveals the presence of Black and multiracial communities in Georgian London. Trial records remain arguably the most detailed source for exploring where, and with whom, ordinary Black people lived, worked, and socialised across the city. This article adapts Kenneth Little’s definition of a community, characterised ‘by a common background of experience’, from his study of race in post-war Cardiff to the context of eighteenth-century London. Communities in the Georgian metropole were formed from a variety of shared experiences, such as heritage dispossession, poverty, work and socialisation, as well as the forced necessity of sharing homes. Black people were part of many, often overlapping, communities of experience, rooted in a contextually heightened desire for friendship, protection and belonging. An understanding of the composition of these communities is integral for reconstructing how ordinary Black people experienced eighteenth-century London
Profiling the LAM Family of Contact Site Tethers Provides Insights into Their Regulation and Function
Full text linkMembrane contact sites are molecular bridges between organelles that are sustained by tethering proteins and enable organelle communication. The endoplasmic reticulum (ER) membrane harbors many distinct families of tether proteins that enable the formation of contacts with all other organelles. One such example is the LAM (Lipid transfer protein Anchored at Membrane contact sites) family in yeast, which is composed of six members, each containing a putative lipid binding and transfer domain and an ER-embedded transmembrane segment. The family is divided into three homologous pairs each unique in their molecular architecture and localization to different ER subdomains. However, what determines the distinct localization of the different LAMs and which specific roles they carry out in each contact are still open questions. To address these, we utilized a labeling approach to profile the proximal protein landscape of the entire family. Focusing on unique, candidate interactors we could support that Lam5 resides at the ER-mitochondria contact site and demonstrate a role for it in sustaining mitochondrial activity. Capturing shared, putative interactors of multiple LAMs, we show how the Lam1/3 and Lam2/4 paralogous pairs could be associated specifically with the plasma membrane. Overall, our work provides new insights into the regulation and function of the LAM family members. More globally it demonstrates how proximity labeling can help identify the shared or unique functions of paralogous proteins
The role of Artificial Intelligence in the construction of news: Challenges and opportunities facing journalism in an age underlined by increasing distrust in knowledge-producing institutions.
Full text linkNews is constructed through a myriad of processes reflecting the cultural and social context in which newsrooms operate as well as the work routines and ownerships structures that govern news organisations. Natural language processing (NLP) and machine learning algorithms have now enabled news organisations to automate content creation, significantly improving efficiency. These algorithms can analyse data, generate headlines, and write news articles. Such innovations have opened opportunities for journalists to focus on investigative journalism and in-depth reporting, while also providing real-time news to an information-hungry audience. However, the rise of AI in news construction also brings its own set of challenges, one of the most significant issues being trust. This paper will discuss how AI is currently used in news organizations, highlighting successful projects and lessons learned. The democratisation of content creation and the potential for personalised, data-driven news experiences also hold immense promise. Yet the industry must grapple with profound issues of trust, ethics, and transparency to maintain the integrity of journalism in an era where traditional knowledge-producing institutions are met with scepticism
Distinguishing mechanisms of social contagion from local network view
Full text linkThe adoption of individual behavioural patterns is largely determined by stimuli arriving from peers via social interactions or from external sources. Based on these influences, individuals are commonly assumed to follow simple or complex adoption rules, inducing social contagion processes. In reality, multiple adoption rules may coexist even within the same social contagion process, introducing additional complexity into the spreading phenomena. Our goal is to understand whether coexisting adoption mechanisms can be distinguished from a microscopic view, at the egocentric network level, without requiring global information about the underlying network, or the unfolding spreading process. We formulate this question as a classification problem, and study it through a likelihood approach and with random forest classifiers in various synthetic and data-driven experiments. This study offers a novel perspective on the observations of propagation processes at the egocentric level and a better understanding of landmark contagion mechanisms from a local view
AI and the Law: assistant or assassin
No full textThe increasing prominence of generative artificial intelligence (GenAI), particularly large language models (LLM), has sparked significant legal discourse, particularly surrounding copyright, patent, and intellectual property laws. In the UK and Europe, AI companies are negotiating licensing deals with media outlets to use human-made content for training purposes. However, a legal tension arises as content owners argue that the unlicensed use of their material amounts to copyright infringement, particularly in the music industry. While ongoing litigation in the US explores whether AI systems can legally scrape copyrighted works to train models, the European approach leans toward alternative dispute resolutions and licensing agreements. A noteworthy case, Emotional Perception AI Ltd, addresses whether artificial neural networks (ANN) can be patented under s 1(2) of the Patents Act 1977. The Court of Appeal ruled an ANN is akin to computer program and therefore does not meet the threshold for patentability, as they lack a ‘technical effect.’ The case is pending appeal before the UK Supreme Court.
GenAI’s impact on creative industries is reshaping legal perspectives on copyright in both the US and the UK. Music AI tools, such as Meta's AudioCraft, are capable of generating original compositions by analysing vast amounts of pre-existing works. This raises complex questions about the ownership of AI-generated music, with potential legislative changes in the UK considering whether AI creators should be granted copyright protection or if existing IP owners should retain control over their work.
Furthermore, the legal profession itself is witnessing the integration of AI into everyday legal practice, from drafting contracts to performing legal research. While AI provides efficiencies, concerns about data protection and confidentiality remain significant, with some jurisdictions introducing specific regulations governing AI’s use in legal proceedings. As generative AI continues to evolve, its legal implications will require ongoing adaptation to balance innovation with the protection of intellectual property and rights holders
Grounded Empiricism
Full text linkEmpiricism has a long and venerable history. Aristotle, the Epicureans, Sextus Empiricus, Bacon, Locke, Hume, Mill, Mach and the Logical Empiricists, among others, represent a long line of historically influential empiricists who, one way or another, placed an emphasis on knowledge gained through the senses. In recent times the most highly articulated and influential edition of empiricism is undoubtedly Bas van Fraassen’s constructive empiricism. Science, according to this view, aims at empirically adequate theories, i.e. theories that save all and only the observable phenomena. Roughly put, something is observable in van Fraassen’s view if members of the human epistemic community can detect it with their unaided senses. Critics have contested this notion, citing, among other reasons, that much of what counts as knowledge for scientists, especially in the natural sciences, concerns things that are detectable only with instruments, i.e. things that are unobservable and hence unknowable by van Fraassen’s lights. The current paper admits the objection’s judiciousness and, in reaction, investigates what gives sensory organs epistemic credibility. It turns out that their credibility can be traced to some principles that are also satisfied by certain instruments. On the basis of this work, a liberalised conception of observability is proposed and defended, along with a closely linked, and accordingly liberalised, conception of empiricism. ‘Grounded observability’ and ‘grounded empiricism’, as we call them, remain true to the spirit of empiricism, but acknowledge that epistemic credibility extends far beyond biological sensory organs to include scientific instruments
