<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Wilson Chua on Medium]]></title>
        <description><![CDATA[Stories by Wilson Chua on Medium]]></description>
        <link>https://medium.com/@wilson-chua?source=rss-729c841d07b7------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*9on6V9t40EWP7htf</url>
            <title>Stories by Wilson Chua on Medium</title>
            <link>https://medium.com/@wilson-chua?source=rss-729c841d07b7------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 23:19:05 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@wilson-chua/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Visualizing # of Vuln GRE IPinIP Devices in Asia Pac]]></title>
            <link>https://wilson-chua.medium.com/visualizing-of-vuln-gre-ipinip-devices-in-asia-pac-3292ece2bb94?source=rss-729c841d07b7------2</link>
            <guid isPermaLink="false">https://medium.com/p/3292ece2bb94</guid>
            <category><![CDATA[firewall-bypass]]></category>
            <category><![CDATA[ip-tunnels]]></category>
            <category><![CDATA[ipinip]]></category>
            <category><![CDATA[gre]]></category>
            <dc:creator><![CDATA[Wilson Chua]]></dc:creator>
            <pubDate>Fri, 04 Sep 2026 00:08:38 GMT</pubDate>
            <atom:updated>2026-09-04T00:08:38.891Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7rptjkMrDBVQcEwn3qcVUA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2ICAb1RSJsh0QtaIM05TsA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*EetduNjecRVxGWt3eUKZ6A.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*i01F2lo0xK_JGuG-Z1qb4Q.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3292ece2bb94" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Hidden DMARC Bypass: When Your Mail Server Turns a Phishing Email Into a Trusted One]]></title>
            <link>https://wilson-chua.medium.com/the-hidden-dmarc-bypass-when-your-mail-server-turns-a-phishing-email-into-a-trusted-one-e2dd303b16ad?source=rss-729c841d07b7------2</link>
            <guid isPermaLink="false">https://medium.com/p/e2dd303b16ad</guid>
            <dc:creator><![CDATA[Wilson Chua]]></dc:creator>
            <pubDate>Thu, 13 Nov 2025 23:53:15 GMT</pubDate>
            <atom:updated>2025-11-13T23:53:15.221Z</atom:updated>
            <content:encoded><![CDATA[<p>We recently received a phishing email that looked like it came from <strong>bnshosting.net</strong>, even though that domain is protected with SPF, DKIM, and DMARC — the three security systems meant to stop email forgery.</p><p>So the big question is: <br><strong><em>How did the fake email still pass all the security checks?</em></strong></p><p>The short answer:<br><strong>Our very own email servers accidentally made the fake email look real during forwarding.</strong></p><p>Here’s the simple explanation.</p><h3>1. The Email Didn’t Come From BNSHosting at All</h3><p>The attacker actually sent the message from <strong>reservation@redfisharuba.com</strong> using a legitimate email-sending service (MailChannels).</p><p>That service allowed the email to pass basic checks (like SPF and DKIM) for <strong>redfisharuba.com</strong> — so at this stage, the message was real, but from the wrong domain.</p><h3>2. The Attacker Changed the Display Name to Look Like BNSHosting</h3><p>The attacker used this forged “From” line:</p><pre>From: Bnshosting Hosting &lt;cPanel@bnshosting.net&gt;</pre><p>Emails show the <strong><em>display name</em></strong> to the user, not the underlying technical sender. Think of it like writing someone else’s name on the return-address of an envelope. The post office doesn’t check the printed name — only the stamp.</p><p>Email works the same way. Security filters don’t validate the <strong>displayed name</strong>. They validate the <strong>hidden technical sender</strong>. So the attacker simply lied in the part humans see.</p><h3>3. The Email Passed Through BNSHosting’s Own Mail Server</h3><p>Here is the critical part.</p><p>The attacker <strong>sent the phishing email <em>to</em> bnshosting.net</strong>, and our server received it normally. Then our server <strong>forwarded the email internally</strong>, and during this process, it did two things:</p><h3>4. Our Server Rewrote the Sender Address (SRS)</h3><p>This is normal when servers forward mail.<br> It ensures SPF (one of the email checks) doesn’t fail.</p><p>But this rewriting made SPF appear valid <strong>for bnshosting.net</strong>, even though the email was originally from redfisharuba.com.</p><h3>5. Our Server Accidentally Added a BNSHosting DKIM Signature</h3><p>This is the biggest mistake. ur server saw:</p><ul><li>The email passed SPF</li><li>The email came through an authenticated relay</li><li>The envelope sender was rewritten to use bnshosting.net</li></ul><p>So our mail system decided:</p><blockquote><em>“This must be one of our own emails. Let’s sign it with our DKIM key.”</em></blockquote><p>When it did that, it added a cryptographic signature saying:<br><strong>“This message really is from bnshosting.net.”</strong></p><p>Even though it wasn’t.</p><p>This completely fooled DMARC, because DMARC trusts DKIM from the domain shown in the “From” line.</p><h3>6. Result: A Fake Email Became Fully ‘Legitimate’</h3><p>After your server re-signed the message:</p><ul><li>SPF looked valid</li><li>DKIM looked valid (because your server signed it)</li><li>DMARC looked valid</li><li>The “From” line matched those signatures</li></ul><p>The message now appeared to be a perfectly authentic bnshosting.net email.</p><p>The attacker didn’t break SPF, DKIM, or DMARC.<br> <strong>Your server unintentionally “blessed” the email.</strong></p><h3>What Caused the Problem?</h3><p>A configuration issue:</p><h3>❌ Our server signs forwarded emails even when the message wasn’t originally from bnshosting.net!</h3><p>It should only sign emails if:</p><ul><li>the sender is a real bnshosting.net user</li><li>the message originates from bnshosting.net</li><li>not when the message is merely passing through the server</li></ul><p>Because the attacker sent the email <em>to</em> bnshosting.net, your server ended up giving it official “bnshosting.net credentials.”</p><h3>How to Prevent This in the Future</h3><h3>✔ Disable DKIM signing for forwarded mail</h3><h3>✔ Require that DKIM signing only happens if the authenticated user belongs to bnshosting.net</h3><h3>✔ Enforce strict DMARC alignment rules</h3><p>(so DKIM from redfisharuba.com cannot be counted as DKIM for bnshosting.net)</p><p>With those fixes, our server will no longer “vouch for” emails that come from outside domains.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e2dd303b16ad" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What Data says about Duterte’s ICC Arrest]]></title>
            <link>https://wilson-chua.medium.com/will-dutertes-icc-arrest-backfire-e619a473a8f0?source=rss-729c841d07b7------2</link>
            <guid isPermaLink="false">https://medium.com/p/e619a473a8f0</guid>
            <dc:creator><![CDATA[Wilson Chua]]></dc:creator>
            <pubDate>Tue, 18 Mar 2025 23:18:08 GMT</pubDate>
            <atom:updated>2025-03-19T09:10:59.625Z</atom:updated>
            <content:encoded><![CDATA[<p>Here is an interesting Qsearch report that tracked the social media sentiments around ex-President Duterte’s arrest. This QSearch report was gathered from publicly available social data from Meta platforms using official APIs, not by scraping web pages.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_soPlcGoDzvaI05hQR7LSg.jpeg" /><figcaption>Facebook Content, Total Three Months 2025 (Left) and Social Sentiment (Right) Mar 7 to Mar 17, 2025</figcaption></figure><h3>Massive Surge in Engagement for March 2025</h3><p>Big data analytics reveals a dramatic spike in online discussions about “Duterte” in March 2025. Social media interactions skyrocketed to <strong>65 million</strong>, a sharp increase from <strong>21 million in February</strong> and <strong>5 million in January</strong>. This represents a <strong>more than twelvefold surge</strong> within just two months, indicating a major event that has captured public attention. (Hint: ICC arrest)</p><h3>Dominance of Positive and Humorous Reactions</h3><p>Analyzing engagement trends, <strong>over 85%</strong> of interactions express positive or humorous sentiments. The most common reactions are:</p><ul><li><strong>Like:</strong> 31 million (35%)</li><li><strong>Love:</strong> 18 million (20%)</li><li><strong>Haha:</strong> 28 million (31%)</li></ul><p>This pattern suggests that most social media users perceive the content around “Duterte” in a favorable or lighthearted manner, highlighting the role of sentiment analysis in understanding public opinion.</p><h3>Presence of Negative Sentiment and Polarization</h3><p>Despite the overall positive engagement, data analytics also detects significant <strong>negative sentiment</strong>:</p><ul><li><strong>Angry:</strong> 7 million (7.78%)</li><li><strong>Sad:</strong> 5 million (6.20%)</li></ul><p>This polarization highlights a divided public response, demonstrating how big data can uncover diverse opinions and gauge societal reactions.</p><h3>The Halo Effect: Political Implications of Public Engagement</h3><p>The surge in engagement may have real-world electoral consequences. A “halo effect” could benefit pro-Duterte candidates in the <strong>May 2025 midterm elections</strong>, as increased visibility and sentiment transfer influence voter behavior.</p><h4>Key Insights from Big Data Analysis:</h4><ol><li><strong>Increased Visibility and Favorability</strong></li></ol><ul><li>The spike in engagement (66.37 million interactions) reflects strong public interest.</li><li>High levels of “Like” (34.41%) and “Haha” (30.90%) reactions suggest Duterte’s enduring popularity.</li><li>This positive perception may extend to candidates affiliated with him, boosting their electoral prospects.</li></ul><p><strong>2. Electoral Boost for Pro-Duterte Candidates</strong> According to another report, this time from Tangere’s <strong>2025 Pre-Election Senatorial Survey that was also </strong>conducted around Duterte’s arrest, notable gains were observed among <strong>PDP-Laban candidates</strong>:</p><ul><li><strong>Senator Bong Go:</strong> 51.25% → <strong>54.00%</strong> (Ranks 1–3)</li><li><strong>Senator Bato Dela Rosa:</strong> 36.54% → <strong>37.29%</strong> (Ranks 7–13)</li><li><strong>Actor Philip Salvador:</strong> 21.92% → <strong>22.67%</strong> (Ranks 19–20)</li><li><strong>Cong. Rodante Marcoleta:</strong> 19.12% → <strong>19.83%</strong> (Ranks 21–23)</li><li><strong>Atty. Jimmy Bondoc:</strong> 17.63% → <strong>18.13%</strong> (Ranks 21–23)</li><li><strong>Atty. Vic Rodriguez:</strong> 6.63% → <strong>10.42%</strong> (Ranks 26–27)</li></ul><p>Additionally, <strong>85% of respondents</strong> in Tangere’s survey confirmed that recent events would influence their voting decisions. However, Roger Do, AutoPolitic cautions that this effect may not hold for long:</p><blockquote>“The initial surge in engagement was merely a reaction to the shock of the arrest, <em>and </em><strong><em>it is already fading</em></strong>.</blockquote><blockquote>The key driver of public interest appears to be the perception that he was previously untouchable. The conversation has shifted towards sovereignty and legal processes, demonstrating how different crisis generate different grades of justification from their campaign, and how the public reacts to them”</blockquote><p><strong>3. Polarization and Mobilization Potential</strong></p><ul><li>While positive reactions dominate, negative sentiment (“Angry” at 7.78% and “Sad” at 6.20%) indicates a divided audience.</li><li>This polarization could <strong>energize both supporters and detractors</strong>, leading to increased voter turnout.</li><li>Pro-Duterte candidates might benefit from the mobilization of his loyal base, particularly in strongholds.</li></ul><p><strong>Historical Influence of Duterte’s Endorsements</strong></p><ul><li>Past elections demonstrate Duterte’s power in shaping voter preferences.</li><li>In 2022, his support was a major factor in <strong>Bongbong Marcos’ victory</strong>.</li><li>Current trends suggest that candidates aligned with Duterte may again see an advantage due to name recall and public sentiment.</li></ul><h3>The Power of Data Analytics in Political Analysis</h3><p>The unprecedented surge in social media engagement surrounding ex-President Duterte’s arrest highlights the power of big data in understanding public sentiment and its potential political consequences.</p><p>In this era where online sentiment can drive real-world political outcomes, understanding and leveraging big data is no longer an option — it’s a necessity.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e619a473a8f0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Getting Better Outcomes: Why 4Ps Should Target Educators, Not Just Students]]></title>
            <link>https://wilson-chua.medium.com/getting-better-outcomes-why-4ps-should-target-educators-not-just-students-0870559268fd?source=rss-729c841d07b7------2</link>
            <guid isPermaLink="false">https://medium.com/p/0870559268fd</guid>
            <category><![CDATA[philippine-education]]></category>
            <category><![CDATA[ps4]]></category>
            <dc:creator><![CDATA[Wilson Chua]]></dc:creator>
            <pubDate>Mon, 27 Jan 2025 22:39:00 GMT</pubDate>
            <atom:updated>2025-01-27T22:39:00.992Z</atom:updated>
            <content:encoded><![CDATA[<p>For decades, conditional cash transfers (CCTs) like the Philippines 4Ps program have been a instrumental in keeping students enrolled in classes. These programs, which provide financial incentives to families conditional on school attendance or performance, have boosted enrollment rates and kept vulnerable children in classrooms. But as global learning crises persist — evidenced by stagnant test scores and widening equity gaps — it’s time to ask: Are we getting the right returns from such investments?</p><p>Emerging research suggests that redirecting CCTs to motivate educators, rather than solely students, could yield transformative results. Here’s why policymakers should consider this paradigm shift.</p><p><strong>The Limits of Student-Focused CCTs</strong> <br>CCTs for households have undeniably improved <strong><em>access</em></strong> to education. Meta-analyses of <a href="https://www.nber.org/papers/w23594">47 programs</a> across low- and middle-income countries confirm their success in raising enrollment and attendance, particularly at the secondary level . Yet these gains often fail to translate into better learning outcomes. For example, Brazil’s “Bolsa Família” increased school participation but showed no significant impact on literacy rates . Similarly, while CCTs reduce dropout rates, their effect on academic achievement is inconsistent, as many programs overlook systemic barriers like teacher quality and classroom resources. In the Philippines, <strong><em>graduation rates have actually dropped over the years.</em></strong></p><p>In short, CCTs address symptoms, not root causes. Without motivated, skilled teachers, even the most diligent students struggle to thrive.</p><p><strong>The Case for Educator-Targeted CCTs</strong></p><p>Shifting incentives to educators could address two critical challenges: teacher retention and instructional quality. Here’s how:</p><p>1. Tackling Teacher Shortages in Underserved Areas <br>Globally, rural and low-income schools face chronic teacher shortages. Career-change teachers (CCTs) — professionals transitioning into education from other fields — often fill these gaps but face steep challenges. Studies show that midcareer entrants struggle to reconcile their prior identities with teaching roles, leading to high attrition rates .</p><p>Conditional cash transfers for educators could incentivize retention. For instance: <br>- Rural retention bonuses: Payments tied to multiyear commitments in underserved schools. <br>- Professional development stipends: Rewards for completing training in trauma-informed pedagogy or STEM instruction .</p><p>China’s rural teacher subsidies, though not formal CCTs, demonstrate this potential. By offering financial incentives, the country reduced urban-rural teacher gaps and improved student retention .</p><p>2. Enhancing Instructional Innovation <br>Career-change teachers bring diverse skills — from engineering to entrepreneurship — that can revitalize classrooms. However, rigid school systems often stifle their creativity. One teacher lamented: Skills from my previous career are seen as irrelevant here. It limits what I can offer” .</p><p>Educator-focused CCTs could reward innovation. For example: <br>- “Out-of-the-box” teaching grants: Incentives for integrating real-world expertise into lesson plans. <br>- Peer leadership bonuses: Payments for mentoring early-career teachers or leading curriculum reforms .</p><p>Such programs align with evidence that teacher identity development — a sense of purpose and autonomy — is key to job satisfaction and effectiveness .</p><p><strong>Designing Effective Educator Incentives </strong><br>To maximize impact, educator CCTs must avoid the pitfalls of past programs:</p><p>1. Balance Inputs and Outcomes: Tie payments to both measurable goals (e.g., student attendance) and professional growth (e.g., training completion) . <br>2. Target Equity Gaps: Prioritize incentives for schools in marginalized regions. <br>3. Foster Collaboration: Pair CCTs with systemic reforms, such as smaller class sizes or technology upgrades .</p><p>The ADB and other institutions have long invested in student-focused CCTs. Yet as the 2024 NBER meta-analysis notes, “programs that ignore teacher motivation risk diminishing returns”. Redirecting even a fraction of CCT funding toward educators could: <br>- Stabilize staffing in crisis-hit schools. <br>- Empower teachers to innovate. <br>- Break the cycle of *schooling without learning*.</p><p>Education systems cannot thrive on access alone. By investing in educators — the architects of learning — we can build classrooms where every student succeeds. Let’s reimagine CCTs as tools not just for enrollment, but for excellence.</p><p>For further reading, explore the <a href="https://www.nber.org/papers/w29758">NBER’s meta-analysis</a> on CCTs and <a href="https://www.mdpi.com/2227-7102/14/8/857">MDPI’s study on career-change teachers</a> .</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0870559268fd" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[“Too Much Security” brought down Philippine EDU sites?]]></title>
            <link>https://wilson-chua.medium.com/too-much-security-brought-down-philippine-edu-sites-5645df553517?source=rss-729c841d07b7------2</link>
            <guid isPermaLink="false">https://medium.com/p/5645df553517</guid>
            <category><![CDATA[dnssec]]></category>
            <category><![CDATA[ddos-attack]]></category>
            <category><![CDATA[philippines]]></category>
            <dc:creator><![CDATA[Wilson Chua]]></dc:creator>
            <pubDate>Sat, 25 Jan 2025 04:20:54 GMT</pubDate>
            <atom:updated>2025-01-25T04:36:08.249Z</atom:updated>
            <content:encoded><![CDATA[<p>On January 24, 2025, netizens report several websites ending in .edu.ph going ‘dark’ (inaccessible). This sudden and widespread outage raised concerns among users. Some thought an army of hackers had successfully breached all of these .edu.ph sites. However, paradoxically, it appears that an effort to <strong><em>enhance</em></strong> security may have inadvertently caused the problem.</p><p><strong>Single Point of Failure: PHNET</strong></p><p>First, the Philippine Educational institutions affected by the outage pointed to PHNET, the registrar responsible for managing the .edu.ph domains, as the likely source of the issue. The situation makes sense because if PHNET service goes down, all edu.ph will be similarly affected. You see, to access a .edu.ph website, users’ devices first need to resolve its domain name through a system called DNS (Domain Name System). PHNET plays a crucial role in this process by directing DNS queries to the correct servers. If this step fails, users encounter a “Not Found” error.</p><p>Once it was established that PHNET was central to the outage, attention turned to what had gone wrong there. Reports from various news sources indicated that a Denial of Service (DoS) attacks targeted PHNET. This type of attack overwhelms servers with excessive requests, making it difficult for legitimate users to access services. PHNET may have overloaded demand. However, some experts expressed uncertainty about this DoS attack without being able to examine and confirm it from the actual logs.</p><p>I attempted to contact Bombim Cadiz (PHNET) for further clarification but did not receive a response before publication. Meanwhile, unnamed sources mentioned that PHNET planned to add more servers to manage the “increased load”.</p><p><strong>DNSSEC Misconfigured or Exploited?</strong></p><p>Secondly, An independent DNS expert, Giancarlo Guiao identified a “broken trust error” during his investigation. He discovered that PHNET incorrectly configured DNSSEC (Domain Name System Security Extensions). This DNSSEC may have backfired and could have cause the outage. See this error below:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6JoKHamMD0gK05PwhO2ITQ.png" /></figure><p>DNSSEC uses cryptographic keys for added security, but issues like expired keys or misconfigurations <strong><em>can disrupt service</em></strong>. In addition, DNSSEC also increases complexity and likewise increases the chances of human error during setup and maintenance. Such failures can prevent users from connecting to websites and applications relying on DNS. Mr Guiao’s further digging showed this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/715/1*vbft8gXbaprbUo2Ip74kUw.png" /><figcaption>Possible misconfiguration of DNSSEC. Screenshot during the outage</figcaption></figure><p>Moreover, there are concerns that DNSSEC itself has vulnerabilities that hackers could exploit. A report from APNIC highlighted a critical flaw known as “KeyTrap,” which can be triggered by a single malicious DNS packet. This flaw can exhaust server resources and disrupt internet access for systems using DNSSEC. <a href="https://blog.apnic.net/2024/02/19/keytrap-algorithmic-complexity-attacks-exploit-fundamental-design-flaw-in-dnssec/">https://blog.apnic.net/2024/02/19/keytrap-algorithmic-complexity-attacks-exploit-fundamental-design-flaw-in-dnssec/</a></p><p>However, it’s essential to note that these observations are speculative until PHNET itself provides an official explanation on the outage. Thankfully, as of this writing, the PHNET is operational and so are the .edu.ph sites using it. Here’s to hoping there won’t be a repeat of this in the future.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5645df553517" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[ML to Detect Spoofed IP Addresses: A Study in Progress]]></title>
            <link>https://wilson-chua.medium.com/ml-to-detect-spoofed-ip-addresses-a-study-in-progress-464ec7d6d29a?source=rss-729c841d07b7------2</link>
            <guid isPermaLink="false">https://medium.com/p/464ec7d6d29a</guid>
            <category><![CDATA[spoofed-ip-detection]]></category>
            <category><![CDATA[big-data-analytics]]></category>
            <category><![CDATA[network-security]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[random-forest]]></category>
            <dc:creator><![CDATA[Wilson Chua]]></dc:creator>
            <pubDate>Fri, 17 Jan 2025 02:06:33 GMT</pubDate>
            <atom:updated>2025-01-17T02:09:53.765Z</atom:updated>
            <content:encoded><![CDATA[<p>Distributed Denial-of-Service (DDoS) attacks have doubled in 5 years (2018 to 2023). It also evolved over time. We see hackers using Spoofed IP addresses while launching attacks from compromised servers. Hackers adopted this tactic of using spoofed IP addresses to “cover their tracks”. Changing the IP addresses makes it appear as though the attacks come from somewhere ELSE. This makes it more difficult to trace and block the attacks.</p><p>The significance of this research becomes apparent when considering the harmful impacts on victims. These include a loss of trust, heightened susceptibility to <em>further</em> attacks, and increased complexity that hinders effective mitigation efforts — implementing robust countermeasures depend on accurately identifying the source IP. The ability to detect spoofed IPs has become an urgent necessity in safeguarding systems against evolving cyber threats.</p><p>A study is currently underway that explores how Machine Learning (ML) can be used to detect spoofed IP addresses. This can identify and classify source IP addresses as either legitimate (“Real”) or maliciously altered (“Spoofed”). Machine Learning offers a scalable, data-driven approach to tackle this issue by analyzing patterns and anomalies that traditional methods might miss. One choice of algorithm is the Random Forest Classification. Early results have been promising so far:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/652/1*g2UvwRyxEwFiDHiB71b4Lg.png" /><figcaption>Confusion Matrix showing performance of Random Forest Classification</figcaption></figure><p>This robust method is also ‘explainable’ algorithm. Random forest provides the reader with the basis for the model. The research on Random Forest models for this purpose marks an important step forward in cybersecurity, equipping organizations with advanced tools to defend against ever-evolving threats in the digital landscape.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=464ec7d6d29a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Solving the Last Mile Challenge: Unconnected.org Brings Internet to the Unserved]]></title>
            <link>https://wilson-chua.medium.com/solving-the-last-mile-challenge-unconnected-org-brings-internet-to-the-unserved-fc352227ecfc?source=rss-729c841d07b7------2</link>
            <guid isPermaLink="false">https://medium.com/p/fc352227ecfc</guid>
            <category><![CDATA[starlink]]></category>
            <category><![CDATA[unconnected]]></category>
            <dc:creator><![CDATA[Wilson Chua]]></dc:creator>
            <pubDate>Tue, 12 Nov 2024 13:32:32 GMT</pubDate>
            <atom:updated>2024-11-12T13:43:49.750Z</atom:updated>
            <content:encoded><![CDATA[<p>Imagine that your school is among the many that is located in a remote rural village in the Philippines. Your students need internet, but your area is “un-served”. This means that there are NO telco facilities. There is no internet connection.</p><p>There were two usual ways to get internet to your village. You can have fiber laid out to your community, or you can use Wireless radio. However, your village/community can’t afford to pay for dedicated fiber to connect your village to the nearest internet hub. It is also blocked by mountains so there is no clear line of sight for radio links. What can you do on a tight budget? You contact Unconnected.org! Cabulisan Elementary School did just that.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/659/1*Q1Mmg-1f_OSq9eV2Hr3Nzw.png" /><figcaption>Photo: <a href="http://unconnected.org/">unconnected.org</a>, USAID BEACON, DICT and Faculty of Cabulisan Elementary School, and Brgy Officials, at Inopacan, Leyte during the activation of community network last August 2024</figcaption></figure><p>Unconnected.org, a visionary non-profit organization, in partnership with <a href="https://blogs.microsoft.com/on-the-issues/2023/05/16/microsoft-airband-latin-america-africa/">Microsoft Airband</a>. Their mission is to connect <strong>1 million Filipinos by 2025</strong>. This initiative will transform education and community connectivity in some of the country’s most isolated areas.</p><p><strong>Bridging the Digital Divide for Rural Schools and Communities</strong></p><p>Unconnected.org works by providing a 3rd way — using Low Earth orbit Satellite internets. They partner with local internet service providers (ISPs) to build community networks using low-earth orbit (LEO) satellite technology. (Aka: Starlinks)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/902/1*BaTIEFnlQDUc_EsO2a932w.png" /><figcaption>Most recent data on Starlink deployments submitted by ProjectBASS.org volunteers</figcaption></figure><p>This cutting-edge model offers FREE internet access to schools, empowering them to bridge the education gap in underserved regions, while enabling ISPs to offer affordable internet to surrounding communities.</p><p>Unconnected.org <strong>eliminates</strong> the need for high startup capital. This lowers the barriers for smaller ISPs, allowing them to sustainably reach more remote areas.</p><p><strong>Prioritizing Schools for a Strong Foundation in Connectivity</strong></p><p>Recognizing the critical role of education, Unconnected.org prioritizes schools in its connectivity program. This foundational access to digital resources opens up new possibilities for students and teachers alike, preparing young learners for a connected future while supporting educators with vital digital tools. Through strategic partnerships, Unconnected.org ensures that schools are well-equipped to lead their communities toward digital empowerment.</p><p><strong>Engaging ISPs and Community Stakeholders for Lasting Impact</strong></p><p>To drive local change, Unconnected.org partners with ISPs, cable operators, and other technology providers to serve regions lacking fiber infrastructure or financial capacity for traditional connectivity solutions. By leveraging strong relationships with satellite operators and network providers, Unconnected.org brings preferential deals and global expertise to local communities, offering access to reliable, education-grade internet solutions.</p><p><strong>Become a Partner in Connectivity</strong></p><p>Schools, ISPs, and companies looking to make a difference in the Philippines have a unique opportunity to support this program. By partnering with Unconnected.org, schools gain free internet access, ISPs expand their reach, and companies explore impactful CSR initiatives — all contributing to the transformation of remote communities.</p><p><strong>Take the Next Step: Connect with Unconnected.org</strong></p><p>Be part of this cause. Reach out today at <a href="/service/mailto:philippines@unconnected.org">philippines@unconnected.org</a>. Build an internet-enabled future that empowers students, educators, and entire communities with the tools they need for growth, learning, and resilience.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fc352227ecfc" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI Detectives: How Machine Learning is Taking on Hackers in the Digital Age]]></title>
            <link>https://wilson-chua.medium.com/ai-detectives-how-machine-learning-is-taking-on-hackers-in-the-digital-age-a79668c0100b?source=rss-729c841d07b7------2</link>
            <guid isPermaLink="false">https://medium.com/p/a79668c0100b</guid>
            <dc:creator><![CDATA[Wilson Chua]]></dc:creator>
            <pubDate>Tue, 05 Nov 2024 12:06:21 GMT</pubDate>
            <atom:updated>2024-11-05T12:06:21.375Z</atom:updated>
            <content:encoded><![CDATA[<p>In the high-stakes world of digital transformation, a groundbreaking study (<a href="https://www.mdpi.com/2227-9709/11/4/83">https://www.mdpi.com/2227-9709/11/4/83</a> ) reveals that artificial intelligence could be pivotal tool in defending e-commerce platforms against hackers.</p><p>With the rise in digital transformation, businesses’ most valuable assets are increasingly centered around their data. However, these vast data troves are magnets (prime targets) for cybercriminals. Among the primary challenges for security specialists today is making sense of enormous weblogs — records of every interaction on a website — to spot the unusual patterns that indicate a security threat. To address this need, researchers behind this study turned to Isolation Forest, a machine learning algorithm designed for anomaly detection, which uses patterns in data to single out suspicious or atypical behavior.</p><p>The research team started by focusing on a publicly available dataset from an e-commerce platform (Iranian e-commerce website <a href="http://zanbil.ir/"><strong>zanbil.ir</strong></a><strong> from Kaggle Dataset)</strong>. Their approach involved a multi-step data preparation pipeline, taking raw weblog data and transforming it into a clean, standardized format suitable for machine learning. The prepared dataset then fed into the Isolation Forest model, which had been trained to identify outliers aka anomalies.</p><p>Isolation Forest is an “unsupervised” algorithm, meaning it doesn’t require labeled examples to learn from — instead, it discovers anomalies on its own. This feature makes it particularly well-suited for cybersecurity, where hackers are constantly evolving their tactics and generating new types of threats that may not be present in existing training data.</p><p>In a head-to-head performance test, the Isolation Forest model demonstrated impressive results, outperforming expert assessments in the detection of suspicious traffic patterns. With an accuracy rate of 93%, the model correctly flagged suspicious activities far more effectively than traditional methods, achieving a 95% precision rate, 90% recall rate, and a robust F1 score of 92%.</p><p>This study used Python’s Scikit-learn, a popular tool in the data science community, to bring this AI approach to life. The Isolation Forest model processed vast data streams at a pace and level of accuracy that could change the game for cybersecurity experts. The researchers believe that with the right data preparation and rigorous model evaluation, Isolation Forest may soon be a staple for companies striving to stay ahead of cyber threats.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/713/1*1-_coyc4jbQeDmlO-CImcg.png" /></figure><p>While the AI system isn’t infallible, its remarkable accuracy suggests that it could become an invaluable partner in cybersecurity, helping specialists monitor and interpret digital activity with unprecedented efficiency. By leveraging machine learning to guard their data, businesses might just have a powerful new ally in the fight against cybercrime.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a79668c0100b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Pinoy Team Shines in SPIRITCYBER-24 Hackathon, Showcasing Pinoy Cybersecurity Talent]]></title>
            <link>https://wilson-chua.medium.com/pinoy-team-shines-in-spiritcyber-24-hackathon-showcasing-pinoy-cybersecurity-talent-c83830e609f6?source=rss-729c841d07b7------2</link>
            <guid isPermaLink="false">https://medium.com/p/c83830e609f6</guid>
            <dc:creator><![CDATA[Wilson Chua]]></dc:creator>
            <pubDate>Thu, 17 Oct 2024 12:15:48 GMT</pubDate>
            <atom:updated>2024-10-17T12:15:48.079Z</atom:updated>
            <content:encoded><![CDATA[<p>In a stellar display of Filipino cybersecurity expertise, Team Peenoise from the Philippines secured 3rd place in the highly competitive SPIRITCYBER-24 IoT/OT Hackathon held during the Singapore International Cyber Week 2024 (Govware). The event, which focused on identifying vulnerabilities in smart city deployments, drew 36 teams from multiple countries, with only 7 advancing to the finals after rigorous remote qualifying rounds.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/587/1*paT5LwqMS4EukDk-VkDgOA.png" /><figcaption>Team Peenoise: Jerold Camacho, Jay Turla and Japz Divino</figcaption></figure><p>Team Peenoise not only placed 3rd but also earned substantial rewards, demonstrating their deep skills in exploiting key vulnerabilities such as Remote Code Execution, XSS, and Modbus attacks on critical IoT/OT systems. Their strong performance resulted in a total prize haul of over $12,550, including the $2,000 third-place bonus.</p><p>The competition featured three challenging bug bounty categories, testing contestants on advanced cybersecurity practices in smart city environments. The hackathon, hosted by the Cyber Security Agency of Singapore and Nanyang Technological University, saw Team Peenoise excel among elite international competitors.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pxD8dcD4LTPPImOjYlH0GA.png" /><figcaption>Qualifying Rounds Scoring</figcaption></figure><p>The top three teams, SeeSeeTV, Salted Egg Chicken, and Peenoise, walked away with impressive bonus prizes of $5,000, $3,000, and $2,000, respectively, for their outstanding performances.</p><p>The SPIRITCYBER-24 hackathon focused on “Cybersecure Urban Living” and invited cybersecurity professionals to demonstrate their talents in defending IoT and OT devices. The competition culminated in a three-day final round in Singapore, where top teams showcased their exploits on real Smart Nation devices.</p><p>Team Peenoise’s success is a testament to the rising prominence of Filipino cybersecurity professionals on the global stage. As one of only seven teams to reach the finals, they exemplified the Philippines’ growing strength in this critical field. Looking forward, the team hopes to return for next year’s challenge, eager to build on their accomplishments and continue representing the Philippines on the international cybersecurity front.</p><p>Maraming salamat to Peenoise: Jay Turla, Japz Divino and Jerold Camacho for their remarkable achievement!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c83830e609f6" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Edward Snowden Warns of AI Power Grabs: Are We Heading Towards a Tech Monopoly?]]></title>
            <link>https://wilson-chua.medium.com/edward-snowden-warns-of-ai-power-grabs-are-we-heading-towards-a-tech-monopoly-e6d82eee06f0?source=rss-729c841d07b7------2</link>
            <guid isPermaLink="false">https://medium.com/p/e6d82eee06f0</guid>
            <dc:creator><![CDATA[Wilson Chua]]></dc:creator>
            <pubDate>Wed, 05 Jun 2024 10:30:57 GMT</pubDate>
            <atom:updated>2024-06-06T01:24:30.594Z</atom:updated>
            <content:encoded><![CDATA[<p>At the recent SuperAI conference in Marina Bay Sands, Singapore, Edward Snowden voiced significant concerns about the evolving landscape of artificial intelligence (AI). He particularly highlighted the increasing efforts to control and monopolize AI development without a thorough understanding of its implications.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/784/1*Ih4HXIpUrM-YccYeinS-jw.png" /><figcaption>Edward Snowden and Ben Goertzel on stage</figcaption></figure><p>Snowden criticized the “home-baked” AI safety initiatives, especially those that inject political agendas into the technology. He pointed to a specific law in California, labeling it flawed and ineffective. He further observed that large corporations often use AI “safety issues” as a pretext to stifle competition from smaller entities. Moreover, Snowden noted that governments leverage the financial resources of companies as a means of exerting control over emerging technologies. He cited the telecommunications industry as an example, where once-dominant players like Nokia and Siemens were eventually overshadowed by Huawei, though he did not mention the U.S. sanctions against Huawei.</p><p>Edward Snowden likened AI to raising a child. Just because a child might grow up to be dangerous doesn’t mean we should stop having children altogether. Remember that most children grow up to be productive members of society while taking care of their elderly. So in the same vein, not raising children because of the fear, is similar to saying that we shouldn’t pursue AI (or AGI) in this case.</p><p>Similarly, wealthy nations and companies are implicitly claiming the right to develop AI while excluding poorer countries and smaller startups. This is like saying only the rich should have the privilege of raising children. This perspective is both unfair and limiting.</p><p>Joining Snowden, Ben Goertzel of Singularity.net remarked on the prevalent fear that advanced AI could inherently pose a threat to humanity. Some believe that an AI, initially beneficial, might turn against humans once it surpasses their intelligence. Snowden countered this by emphasizing the value of cooperation, asserting that AI and humans can achieve greater outcomes together.</p><p>Both Snowden and Goertzel agreed that imposing restrictive laws could hinder AI progress. Goertzel suggested decentralizing AGI deployment, potentially through blockchain technology, as a strategy to avoid regulatory capture and ensure more equitable development of AI.</p><p>Note: <br>Edward Joseph Snowden, born on June 21, 1983, is an American and naturalized Russian citizen. He is famous for leaking highly classified information from the National Security Agency (NSA) in 2013. His disclosures uncovered extensive global surveillance programs run by the NSA and the Five Eyes intelligence alliance, involving cooperation with telecommunication companies and European governments. This ignited a major discussion on national security and individual privacy.</p><p>Snowden worked for Booz Allen Hamilton, Dell, and the CIA, and grew disillusioned with the surveillance programs. His attempts to address ethical concerns internally were ignored. He is currently indicted for espionage and living in exile in Russia. Despite controversy, his revelations have led to significant legal and policy changes regarding privacy and mass surveillance in the U.S.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e6d82eee06f0" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>