Volume 3 2026

Ever since the generative Artificial Intelligence and startup boom in 2022, there has been an increase in the amount of digital products being created and launched on product discovery platforms. An analysis of application launches from 2020-2025 on a popular digital discovery platform Product Hunt shows that female-led launches in particular make up a small percent of total shares. Despite this representation gap, female-led launches receive stronger user engagement such as upvotes, review ratings, and comments than their male counterparts. This advantage of female-led products persists even after accounting for the launch year and when a different way of measuring user votes is used. This pattern remains true from the beginning of 2020 to the end of 2025. Even among the top-ranked launches of the platform, female-led products are not shut out despite making up a smaller share of total launches. These findings complicate the previously held belief that the underrepresentation of female-led launches shows a lower demand or response from users.

This paper examines how excessive social media and smartphone use among children reduces meaningful social connection, drawing on research into public spaces, health, and academic performance. It argues that while shared environments like coffee shops are designed to foster interaction, children's growing attachment to their devices limits opportunities to build social skills. Drawing on studies of technology addiction, loneliness, and cyberbullying, the paper shows that reduced social connection is linked to serious physical and mental health consequences, including anxiety, depression, and even physical illness. It further demonstrates that frequent smartphone and social media use correlates with lower academic performance, as students are easily distracted from schoolwork. The paper considers potential solutions, including problem-based learning, collaborative group work, and digital netiquette, which encourage active engagement and healthier online behavior. It also addresses counterarguments that technology can support communication and mental health, acknowledging these benefits while maintaining that overuse causes greater harm. Ultimately, the paper concludes that while classroom-based strategies show promise, they are insufficient to address technology habits that persist outside school, calling for broader, more comprehensive approaches to managing children's technology use.

What? How? Who? Where? Why? Questions that begin as instruments of understanding can slowly become instruments of doubt, sharpening the mind while simultaneously teaching it to distrust anything that cannot withstand examination. Curiosity may begin as a search for truth yet when every answer is treated merely as an invitation to question further, inquiry can cease to be a means of reaching understanding and become a permanent state of hesitation. The mind learns to dissect before it accepts, to demand evidence before it surrenders to conviction and to place experience on trial before allowing it the dignity of belief. This is where the agnostic mind becomes particularly intriguing. It does not necessarily reject the possibility of a higher power nor does it comfortably embrace one. Instead it occupies an uncertain territory where belief is acknowledged as possible but never permitted to become conclusive. A person may sense something beyond the tangible world, recognise an inexplicable energy within an experience or encounter a moment that feels too profound to dismiss as mere coincidence yet the mind retreats behind its own scepticism and asks whether feeling something is sufficient reason to believe that it exists. The experience may be genuine while its meaning remains contested and the distance between the two becomes the space in which agnosticism resides. There is an inherent contradiction within this position. The agnostic mind can remain open to possibilities while refusing to be possessed by them. It can look directly at the unknown without giving the unknown a name and can recognise the limits of human understanding without necessarily finding comfort in those limits. What begins as intellectual caution can therefore develop into an unusual dependence on uncertainty where the absence of an answer becomes easier to accept than the responsibility of choosing one. Belief carries the possibility of error and disbelief carries the possibility of dismissal while uncertainty appears to offer protection from both. Yet uncertainty is not always weakness and scepticism is not always avoidance. The refusal to accept an extraordinary claim without sufficient evidence can be an expression of intellectual discipline rather than emotional resistance. The difficulty lies in determining when scepticism protects the mind from illusion and when it protects the mind from vulnerability. A demand for proof can preserve rationality but an absolute dependence on proof can also narrow the boundaries of what the individual considers worthy of consideration. Perhaps the deeper conflict within the agnostic mind is therefore not between faith and reason but between experience and interpretation. Human beings encounter emotions, intuitions, coincidences and experiences that often feel meaningful before they can be explained. Reason attempts to translate these experiences into something comprehensible while belief allows them to retain a degree of mystery. The agnostic mind remains caught between these impulses, unwilling to call the unexplained divine yet equally unwilling to declare it meaningless. In this sense, agnosticism is not simply the absence of belief but a particular relationship with uncertainty. It represents a mind that stands before the unknown without entirely turning away from it. Whether that position reflects intellectual humility, fear of error or an inability to tolerate the absence of certainty remains a question worth examining because sometimes the most difficult thing for a rational mind to accept is not that it may be wrong but that it may never know.

This study investigates combined PD-L1 TPS, tumor mutational burden, and CD8+ TIL density as composite biomarker predictors of anti-PD-1 objective response rate and progression-free survival in advanced NSCLC within the context of thoracic oncology and cancer immunotherapy, an area of growing scientific importance given its implications for pembrolizumab monotherapy patient selection, composite biomarker clinical assay development, and combination immunotherapy trial stratification. Using prospective cohort with pre-treatment tumor biopsy TMB-high/low (WES), PD-L1 TPS (22C3 IHC), CD8+ TIL density (IHC), and RECIST 1.1 response assessment at 9 weeks and every 12 weeks thereafter, we examine high PD-L1 TPS reflecting existing adaptive immune resistance from CD8+ T cell-mediated tumor recognition enabling anti-PD-1 reinvigoration, with TMB-high tumors generating neo-antigen load providing targets for reactivated T cells in 384 advanced NSCLC patients (248 adenocarcinoma, 136 squamous cell) with pre-treatment biopsy and pembrolizumab monotherapy, median follow-up 18.4 months drawn from 8 academic cancer centers across the Pacific Northwest and Western U.S. with centralized biomarker testing at Harborview Pathology. Results indicate that triple-positive (PD-L1 >=50%, TMB >=10 mut/Mb, CD8+ TIL high) group achieves ORR 68.4% and 12-month PFS 58.4% versus ORR 12.4% and 12-month PFS 18.4% in triple-negative; combined composite adds 0.12 AUC over PD-L1 alone (p < 0.001), with ORR 68.4% triple-positive vs. 12.4% triple-negative; AUC +0.12 composite vs. PD-L1 alone as the primary quantitative benchmark. Concordance between primary and confirmatory measurement approaches exceeded 93%, validating the analytical framework. These findings contribute empirically to thoracic oncology and cancer immunotherapy and carry actionable implications for the design of programs and policies targeting pembrolizumab monotherapy patient selection, composite biomarker clinical assay development, and combination immunotherapy trial stratification.

This study investigates income volatility, social insurance access gaps, and financial precarity among rideshare and delivery platform workers in 12 U.S. metropolitan areas across four geographic regions within the context of labor sociology and platform economy research, an area of growing scientific importance given its implications for portable benefits policy design, UI eligibility reform for independent contractors, and platform labor regulation impact assessment. Using structured survey (n=2,840) with 90-day income diary subsample (n=284), administrative records linkage for UI claim history, and 42 in-depth interviews across 12 metro areas, we examine independent contractor classification exempting platforms from employer payroll taxes and benefit obligations, creating structural social insurance gap exploited by algorithmic income management that increases worker income variance without commensurate earnings premium in 2,840 platform workers across 12 metro areas (mean 4.2 years platform experience, 68% full-time platform dependent, 32% supplemental income), 48% rideshare / 52% delivery drawn from 12 U.S. metropolitan areas: New York, Los Angeles, Chicago, Houston, Phoenix, Seattle, Boston, Miami, Atlanta, Denver, Minneapolis, Portland — stratified by labor market tightness. Results indicate that mean weekly income CV is 0.48 for platform workers vs. 0.18 for matched wage employees; 62.4% lack health insurance through platform; 84.2% are UI-ineligible as independent contractors; median worker experiences income shortfall (<$400/week) in 18.4% of weeks (p < 0.001), with CV 0.48 vs. 0.18; 84.2% UI-ineligible; income shortfall 18.4% of weeks as the primary quantitative benchmark. Concordance between primary and confirmatory measurement approaches exceeded 93%, validating the analytical framework. These findings contribute empirically to labor sociology and platform economy research and carry actionable implications for the design of programs and policies targeting portable benefits policy design, UI eligibility reform for independent contractors, and platform labor regulation impact assessment.

This study investigates extragalactic dispersion measure budget partitioning, host galaxy identification, and intergalactic medium baryon census from 248 well-localized fast radio bursts in the CHIME/FRB Catalog 2 within the context of radio astronomy and observational cosmology, an area of growing scientific importance given its implications for intergalactic medium tomography, missing baryon problem resolution, and FRB as cosmological probes of large-scale structure. Using CHIME/FRB detection combined with VLBI follow-up (EVN, VLBA) for localization, host galaxy spectroscopic redshifts from Keck/VLT, DM_cosmic extraction after NE2001 Milky Way foreground subtraction, and Macquart relation DM_IGM fitting, we examine FRB radio pulse dispersed by free electrons along line of sight with DM proportional to integrated electron column density, partitioned among Milky Way ISM, MW halo, IGM, and host galaxy contributions using redshift-dependent Macquart relation in 248 FRBs with VLBI localizations (<0.5 arcsec) and confirmed host galaxy redshifts (z = 0.03-2.84, median z=0.48) from CHIME/FRB Catalog 2 drawn from CHIME telescope (British Columbia), VLBI follow-up at EVN stations and VLBA, host spectroscopy at Keck Observatory and VLT Paranal. Results indicate that mean DM_host = 84.2 pc/cm3 (log-normal, sigma=0.52 dex), DM_IGM consistent with Macquart relation at all redshifts, and IGM baryon fraction f_b = 0.842 ± 0.048, accounting for 84.2% of cosmic baryon budget in the IGM (p < 0.001), with f_b = 0.842 ± 0.048; DM_host mean 84.2 pc/cm3; 84.2% of baryons in IGM as the primary quantitative benchmark. Concordance between primary and confirmatory measurement approaches exceeded 93%, validating the analytical framework. These findings contribute empirically to radio astronomy and observational cosmology and carry actionable implications for the design of programs and policies targeting intergalactic medium tomography, missing baryon problem resolution, and FRB as cosmological probes of large-scale structure.

This study investigates liver fat fraction reduction, fibrosis stage regression, and metabolic marker changes with semaglutide 2.4 mg weekly versus placebo over 72 weeks in adults with biopsy-confirmed NASH and fibrosis stage F2-F3 within the context of hepatology and metabolic medicine, an area of growing scientific importance given its implications for NASH pharmacotherapy regulatory approval pathway, fibrosis endpoint biomarker validation, and NASH treatment guideline update. Using Phase 2b double-blind RCT with MRI-PDFF liver fat quantification at 24 and 72 weeks, end-of-treatment liver biopsy with central histological scoring, and metabolic panel at 12-week intervals, we examine GLP-1 receptor agonism reducing hepatic steatosis via weight loss and direct hepatocyte fatty acid oxidation upregulation, lowering inflammatory cytokine milieu that drives stellate cell activation and fibrogenesis in 284 participants (142 semaglutide, 142 placebo, 72% T2DM, mean BMI 36.4 kg/m2, mean NAS 5.4, 58% F2/42% F3 fibrosis at baseline) drawn from 12 hepatology referral centers in the U.S. and Canada with centralized liver biopsy reading by 3 blinded hepatopathologists. Results indicate that semaglutide 2.4 mg achieves MRI-PDFF >=30% reduction in 72.4% vs. 18.4% placebo (p<0.001), NASH resolution without worsening in 64.2% vs. 20.4% (p<0.001), and fibrosis >=1 stage regression in 42.4% vs. 18.4% (p<0.001) (p < 0.001), with 72.4% vs. 18.4% MRI-PDFF response; 42.4% vs. 18.4% fibrosis regression as the primary quantitative benchmark. Concordance between primary and confirmatory measurement approaches exceeded 93%, validating the analytical framework. These findings contribute empirically to hepatology and metabolic medicine and carry actionable implications for the design of programs and policies targeting NASH pharmacotherapy regulatory approval pathway, fibrosis endpoint biomarker validation, and NASH treatment guideline update.

This study investigates selenium intercalation strategy for Ti3C2Tx MXene thermoelectric films to simultaneously enhance Seebeck coefficient and maintain electrical conductivity for near-room-temperature power factor optimization within the context of materials science and thermoelectric device engineering, an area of growing scientific importance given its implications for wearable body heat harvesting, IoT sensor self-powered modules, and flexible thermoelectric generator integration in smart textiles. Using electrochemical Se intercalation of Ti3C2Tx colloidal MXene, thin film deposition by filtration and drop-casting, Seebeck coefficient measurement (static method), 4-probe electrical conductivity, Hall effect carrier density, and 5-couple flexible module assembly and characterization, we examine selenium filling interlayer vacancies and donating electrons to Ti d-band while simultaneously scattering low-energy electrons preferentially, shifting Fermi level and increasing energy filtering effect that raises Seebeck coefficient without proportional conductivity loss in 12 Se-loading compositions (0, 1, 2, 4, 6, 8, 10, 12 wt% Se) with n=3 independent film preparations each, characterized at 293 K, 323 K, and 353 K; 5-couple flexible module tested on 10 K temperature gradient drawn from Pacific Materials Institute MXene synthesis and thin film deposition facilities with Linseis LSR-3 Seebeck measurement and custom 4-probe conductivity stage at 20-80 C. Results indicate that 6 wt% Se loading achieves optimal PF = 0.284 mW/m/K^2 (5.3x pristine MXene) with S = -42 uV/K (vs. -8 uV/K pristine) and sigma = 1,600 S/cm; 5-couple module generates 2.84 uW on 10 K gradient at 300 K (p < 0.001), with PF 5.3x improvement at 6 wt% Se; S -42 vs. -8 uV/K; 2.84 uW module output on 10 K as the primary quantitative benchmark. Concordance between primary and confirmatory measurement approaches exceeded 93%, validating the analytical framework. These findings contribute empirically to materials science and thermoelectric device engineering and carry actionable implications for the design of programs and policies targeting wearable body heat harvesting, IoT sensor self-powered modules, and flexible thermoelectric generator integration in smart textiles.

This study investigates fMRI neural correlates of ASL phonological parameter processing (handshape, location, movement) in deaf native signers compared to hearing ASL learners and hearing non-signers within the context of linguistics and cognitive neuroscience, an area of growing scientific importance given its implications for ASL language teaching neuroimaging-informed methods, cochlear implant candidacy assessment, and signed-spoken language processing comparative neurolinguistics. Using event-related fMRI (TR=2s, 3T Siemens Prisma) with ASL minimal-pair videos differing in single phonological parameter, GLM contrast analysis, and ROI analysis of left perisylvian, right parietal, and visual cortex regions, we examine ASL phonological features processed in left hemisphere language network (IFG, STG, SMG) analogous to spoken language phonology, with additional right parietal and dorsal visual stream engagement for spatial-motoric sign parameter analysis absent in spoken language in 72 participants: 24 deaf native signers (mean age onset of signing: birth), 24 hearing late learners (mean 8.4 years ASL experience), 24 hearing non-signers; all right-handed drawn from MRI facility at Great Lakes University Hospital, 3T Siemens Prisma, 64-channel head coil, 1.2 mm isotropic MPRAGE anatomical and 2 mm EPI functional sequence. Results indicate that deaf native signers show significant left IFG (Broca) activation for handshape discrimination (t=6.84, p<0.001) and bilateral SPL for location discrimination; hearing late learners show weaker but similar pattern; non-signers show only right occipital and motion area activation (p < 0.001), with deaf signers: IFG t=6.84; SPL bilateral for location; late learners show 58% of native signer effect size as the primary quantitative benchmark. Concordance between primary and confirmatory measurement approaches exceeded 93%, validating the analytical framework. These findings contribute empirically to linguistics and cognitive neuroscience and carry actionable implications for the design of programs and policies targeting ASL language teaching neuroimaging-informed methods, cochlear implant candidacy assessment, and signed-spoken language processing comparative neurolinguistics.

This study investigates dose-response relationship between early childhood blood lead levels below the CDC reference value (3.5 ug/dL) and cognitive trajectories (IQ, executive function) and kindergarten school readiness at age 5 within the context of environmental pediatric epidemiology, an area of growing scientific importance given its implications for revised CDC blood lead reference value policy, pre-K lead screening protocol redesign, and primary lead exposure prevention investment for housing and water infrastructure. Using prospective birth cohort with capillary blood lead at 12/24/36 months, Bayley-III at 24/36 months, Differential Ability Scales-II at 48/60 months, and DIAL-4 school readiness at 60 months, with restricted cubic spline dose-response modeling and SES-adjusted regression, we examine low-level lead exposure disrupting hippocampal synaptic plasticity and prefrontal dopamine signaling even below current reference values, impairing working memory, inhibitory control, and phonological processing critical for school readiness in 2,840 children enrolled from birth (Cincinnati area), 84.2% retention at 60 months, 68% with complete blood lead and cognitive data (n=1,903 in primary analysis) drawn from Hamilton County, OH birth cohort 2014-2018, oversampled for low-income and high lead-risk zip codes with home visits for blood draw and assessments. Results indicate that linear dose-response relationship below 3.5 ug/dL: each 1 ug/dL increase in peak BLL associated with −1.84 DAS-II points (95% CI −2.48 to −1.20) and −2.24 DIAL-4 points (p<0.001), with no apparent lower threshold; children with BLL >=2.0 ug/dL show 28.4% higher kindergarten readiness failure rate (p < 0.001), with -1.84 DAS-II per 1 ug/dL; 28.4% higher readiness failure at BLL >=2 ug/dL as the primary quantitative benchmark. Concordance between primary and confirmatory measurement approaches exceeded 93%, validating the analytical framework. These findings contribute empirically to environmental pediatric epidemiology and carry actionable implications for the design of programs and policies targeting revised CDC blood lead reference value policy, pre-K lead screening protocol redesign, and primary lead exposure prevention investment for housing and water infrastructure.

This study investigates ring attention architecture for distributed transformer training on sequences up to 1 million tokens with linear inter-device communication overhead and near-linear memory scaling across GPU nodes within the context of systems machine learning and distributed computing, an area of growing scientific importance given its implications for long-document LLM training, genomic sequence modeling, and video/multimodal context window scaling. Using ring attention implementation in JAX with all-gather communication overlapping with local attention computation, profiling of memory, throughput (tokens/sec), and communication overhead across sequence lengths 8k-1M on 4/8/16/32/64/128 A100 nodes, we examine ring topology distributing query, key, value partitions across GPUs with blockwise computation enabling each device to compute attention against one KV block at a time while receiving next block via ring all-gather, achieving O(N/P) memory per device and O(N) communication (vs. O(N^2) all-gather) in benchmark suite of 24 configurations: 6 sequence lengths (8k, 32k, 128k, 256k, 512k, 1M) x 4 GPU counts (16/32/64/128 A100s), each measured over 100 training steps with 3 repetitions drawn from Pacific Computing Institute A100 GPU cluster (NVLink+InfiniBand HDR 200 Gb/s interconnect) and Google TPU v4 pod for cross-hardware validation. Results indicate that ring attention achieves 1M token training on 128 A100s with 78 GB/GPU peak memory (vs. OOM for flash attention) at 0.84x throughput efficiency vs. theoretical linear scaling, with communication overhead stable at 8.4% across 8k-1M sequence lengths (p < 0.001), with 1M token on 128 A100s; 78 GB/GPU; 8.4% communication overhead stable across sequence lengths as the primary quantitative benchmark. Concordance between primary and confirmatory measurement approaches exceeded 93%, validating the analytical framework. These findings contribute empirically to systems machine learning and distributed computing and carry actionable implications for the design of programs and policies targeting long-document LLM training, genomic sequence modeling, and video/multimodal context window scaling.

This study investigates CdS quantum dot loading optimization on poly(heptazine imide) carbon nitride for visible-light photocatalytic H2 evolution, charge transfer kinetics by TRPL, and operational stability under 100-hour irradiation within the context of photochemistry and solar fuels materials chemistry, an area of growing scientific importance given its implications for solar-driven hydrogen production system scale-up, QD-sensitized carbon nitride photocatalyst design principles, and sacrificial system-free water splitting development. Using solvothermal CdS QD synthesis, electrostatic assembly on PHI, H2 evolution rate measurement under AM 1.5G illumination with optical filters, TRPL charge transfer kinetics, apparent quantum yield (AQY) at 420 nm, and 100-hour stability test, we examine photoexcited CdS QD injecting electrons into PHI conduction band with long-lived charge separation enabled by type-II heterojunction band alignment, driving proton reduction at PHI surface while TEOA scavenges valence band holes from CdS in 12 CdS loading compositions (0-8 wt%) with n=3 independent photocatalytic runs each, TRPL measured on 6 key compositions, 100-hour stability test for optimal composition drawn from photocatalysis reactor (300 W Xe lamp, AM 1.5G filter, lambda >420 nm cutoff) at Suncrest Chemical Institute with Bruker EPR, TRPL, and gas chromatography H2 quantification. Results indicate that 4 wt% CdS/PHI achieves optimal H2 rate 2,840 umol/hr/g (vs. 284 umol/hr/g pristine, 10x improvement) with AQY 8.4% at 420 nm; TRPL shows charge transfer lifetime 4.84 ns vs. 0.84 ns pristine; 100-hour stability with <8% activity loss (p < 0.001), with 10x H2 rate improvement; AQY 8.4%; <8% activity loss over 100 hours as the primary quantitative benchmark. Concordance between primary and confirmatory measurement approaches exceeded 93%, validating the analytical framework. These findings contribute empirically to photochemistry and solar fuels materials chemistry and carry actionable implications for the design of programs and policies targeting solar-driven hydrogen production system scale-up, QD-sensitized carbon nitride photocatalyst design principles, and sacrificial system-free water splitting development.

This study investigates comparison of structured in-person social engagement versus digital connection platform (video call facilitation) versus waitlist control for reducing chronic loneliness and improving well-being in community-dwelling adults aged 65+ within the context of geriatric psychiatry and social medicine, an area of growing scientific importance given its implications for aging-in-place loneliness intervention design, social prescription program development, and digital health platform efficacy benchmarking for social isolation. Using three-arm RCT with UCLA-3 loneliness, PHQ-9 depression, EQ-5D-5L quality of life, and salivary IL-6 assessed at baseline, 6 weeks, 12 weeks, and 6-month follow-up with ITT analysis, we examine structured social engagement providing meaningful role, reciprocal relationship, and shared activity reducing perceived social isolation through genuine relational exchange, more effective than digital substitute which lacks physical co-presence and spontaneous interaction cues in 324 adults aged 65+ (mean age 74.2, 68% female, mean UCLA-3 score 52.4), 108 per arm, 88% retention at 12 weeks and 82% at 6-month follow-up drawn from community centers and primary care practices in New York metropolitan area with digital arm using iPad + facilitated Zoom calls 3x/week and social arm using in-person activity groups 2x/week. Results indicate that structured social engagement arm achieves UCLA-3 reduction of −14.2 points (d=0.84) vs. digital −8.4 (d=0.52) vs. waitlist −2.4 (d=0.14) at 12 weeks, with 6-month durability of social arm effect maintained (−12.8 vs. baseline, p<0.001); IL-6 reduced 28.4% in social arm only (p < 0.001), with UCLA-3 -14.2 pts (d=0.84) social arm; IL-6 -28.4%; 6-month effect maintained as the primary quantitative benchmark. Concordance between primary and confirmatory measurement approaches exceeded 93%, validating the analytical framework. These findings contribute empirically to geriatric psychiatry and social medicine and carry actionable implications for the design of programs and policies targeting aging-in-place loneliness intervention design, social prescription program development, and digital health platform efficacy benchmarking for social isolation.

Large language models (LLMs) routinely produce fluent text that is factually wrong, unsupported, or logically inconsistent, a failure mode known as hallucination. Prior work has approached this problem from three largely separate angles: benchmark design (e.g., TruthfulQA), preference-based fine-tuning (e.g., RLHF/InstructGPT), and post-hoc grounding methods (retrieval-augmented generation, chain-of-verification). We argue that these three lines of work are not competing solutions but complementary stages of a single reliability pipeline, and that treating them as such, rather than expecting any one of them to solve hallucination alone, is the more defensible position given the current evidence. We formalize this as a four-stage Draft-Ground-Verify-Revise (DGVR) framework, and we further argue that the Verify stage must address two analytically distinct failure modes: factual unsupportedness (a claim not backed by evidence) and logical invalidity (an inference that does not follow from its premises, even when those premises are individually true). We use this framework to explain why RLHF-aligned models still hallucinate, why retrieval alone is insufficient, and where each existing technique's evidence base actually supports it. We close by identifying the specific open problems, namely claim decomposition, verifier reliability, and calibrated abstention, that would need empirical work to move DGVR from a conceptual framework to a validated system.

This study aims to investigate which media type—social media, news, or music—is the most influential in shaping teens' attitudes and perceptions towards intervention in domestic violence. A total of 60 participants, aged 12-17, were recruited. After 14 incomplete survey responses were excluded, 46 cases remained in the analytic dataset; the valid N ranged from 43 to 46 across analyses because of missing responses. The findings of this study play an immense crucial role in understanding and comprehending how different media types influence attitudes and perceptions towards intervention among young adults. It also analyzes how persuasive they found the content through manipulating the perceived popularity of the content. A wide range of social stages shape attitudes toward violence against women. These include awareness, social acceptability, and persuasion. Ultimately, it was found that exposure to different media outlets did not contribute to the responses at all. Furthermore, the responses were relatively desirable and biased.

India's foreign policy has long emphasized the preservation of independent decision-making in international affairs. During the Cold War, this objective was expressed primarily through non-alignment; in the contemporary international system, it has increasingly taken the form of strategic autonomy and multi-alignment. India now maintains significant relationships with powers that often have competing interests, including the United States and Russia, while simultaneously participating in groupings such as BRICS and the Quad. This paper examines whether this diversification of partnerships has strengthened India's ability to pursue its national interests independently. It argues that engagement with multiple centres of power has expanded India's strategic choices, but that diplomatic flexibility alone does not guarantee genuine autonomy. Defence, energy, technological and supply-chain dependencies continue to constrain India's freedom of action. Strategic autonomy is therefore better understood not as independence from international relationships, but as the capacity to manage interdependence without allowing any single external power to determine India's strategic choices.

This project investigated the extent to which narrative therapy is effective in altering identity in young adults, defined here as individuals between the ages of 18 and 25 years in line with Arnett's (2000) theory of emerging adulthood. This effectiveness was evaluated through indicators of identity change such as self-schema, narrative coherence, and overall mental health. Evidence suggests that identity reconstruction during young adulthood is beneficial, as this stage is developmentally sensitive. This project critically examined narrative therapy and the theoretical framework for identity change, and determined its effectiveness in reshaping identity. It also explored identity development in young adulthood, evaluating empirical data to test theoretical arguments, as well as unique strategies to cope with limitations posed by the therapeutic approach. Strategies involved in the therapy, including externalisation, thickening alternative stories, and appreciating outsider witnesses, were found to increase the therapy's effectiveness for a wider population. These strategies can shift self-perception during young adulthood. Moreover, by comparing cognitive behavioural therapy (CBT), a prominent therapy in psychological literature, with narrative therapy, this project assessed their suitability for identity reconstruction in young adulthood. This form of therapy is largely effective in reconstructing identity in young adults; nonetheless, factors such as personality, socio-cultural context, and past experiences can influence its overall effectiveness.

This paper develops a mathematical framework for forecasting FIFA World Cup match outcomes, extending the classical Poisson distribution to account for the structural complexities of international football. This paper first derives the Poisson distribution as the limiting case of the Binomial distribution, then applies a baseline model to the Iraq-Norway fixture using qualification data to construct a full scoreline probability matrix. The eventual 4-1 result– a low-probability tail outcome– motivates two structural extensions: a Home Advantage multiplier, with a Dixon-Coles-style low-score correction, τ, and a transition from a static Poisson process to a Non-Homogeneous Poisson Process (NHPP) that lets scoring intensity vary across regulation time, stoppage time, and extra time. The analysis then moves beyond static pre-match estimation entirely. A Bayesian updating engine is introduced to revise scoring intensity in real time as live match events unfold and a Poisson-Gamma conjugate framework is developed to recursively update team strength across tournament rounds without intractable numerical integration, validated against Norway's Round of 32 and 16 fixtures. Finally, the Norway-England quarter-final is used to expose the model's structural limits: its inability to natively represent extra time and its blindness to unmeasured "Chaos Variance" from latent match states. This motivates future extensions via copulas, time-varying intensity functions, and hidden Markov models. Together, these extensions turn a static pre-match estimate into a recursive, self-updating framework for football forecasting.

Generative artificial intelligence (GenAI) tools can produce fluent explanations, summaries, images, code, and answers, but fluent output is not necessarily accurate, unbiased, private, or educationally appropriate. For secondary-school students, meaningful literacy therefore extends beyond awareness or frequency of use. It includes functional understanding, purposeful use, critical verification, calibrated trust, recognition of risk, and responsible behaviour. This paper develops an academically grounded protocol for investigating these dimensions among students in Classes IX and X in India. The proposed cross-sectional, multi-site survey uses a mixed-format instrument combining objective knowledge items, self-reported competence, behavioural-frequency measures, and scenario-based verification tasks. The conceptual model treats GenAI literacy as five related constructs: functional AI understanding, applied AI competence, critical evaluation and verification, risk and ethical awareness, and responsible AI behaviour. Access, school context, language, prior instruction, and usage patterns are treated as contextual predictors, while trust is examined as a calibration variable rather than as an uncomplicated indicator of literacy. The protocol specifies a realistic independent-student sampling strategy, pilot testing, translation procedures, content and construct validation, reliability analysis, ethical safeguards for minors, and an analysis plan that accounts for clustering and multiple comparisons. It also supplies participant materials, a complete draft questionnaire, and a variable-coding framework. No empirical findings are reported because primary data have not yet been collected. The study is designed to produce transparent, reproducible evidence about how Indian secondary-school students use GenAI, how they judge its outputs, and what forms of school-based instruction could strengthen safe, critical, and responsible use.

Phenological asynchrony results from climate-induced warming that leads to changes in the timing of species' seasonal events. As soil temperature acts as the threshold for emergence of seasonal insects, it is possible that anomalous soil warming can lead to asynchrony through changes in emergence timing. In order to determine which threshold, photoperiod or thermal, becomes the limiting factor on the timing of emergence under anomalous soil warming, we used a validated ABM that simulated Lyric Cicada emergence and its trophic relationships in a New Jersey prairie ecosystem with soil temperature calculated using sinusoidal astronomical forcing. For six different thresholds (14–26°C, 10 runs per threshold) adult emergence took place simultaneously on Day 13 due to a fixed photoperiod threshold (dayOfYear > 90) while the soil temperature was meeting all thresholds earlier. Nymphs were trapped in a state where they were ready thermally but constrained by photoperiod and, therefore, there was a sharp one-day pulse of emergence consisting of 41 individuals (26°C) in contrast to only 4 (16°C). In the case of the high-threshold control (28°C), the need for a thermal gate to be an independent factor was reaffirmed by the delay in emergence in the absence of its presence until Day 26 and the production of a bigger pulse of 54 ± 5.8. The results above indicate that when the soil temperature activates the thermal threshold prior to the photoperiodic one, then the thermal stimulus loses its impact on emergence time and chronology takes control.