Merge pull request #156 from arm-university/c_page_changes

Update Challenge_Page.html

Author: matt-cossins

docs/Challenge_Page.html

@@ -198,6 +198,11 @@ <h2>Support & Queries</h2>
           clearly mentioning the problem statement number in the subject line.
         </p>
 
+       <p>For get started resources:</p>
+       <p>
+          <a href="https://arm-university.github.io/online-resources-arm/">Arm Education Resources</a>
+       </p>
+
         <p class="ch-small">All updates will be posted on this page.</p>
      </section>
 
@@ -380,15 +385,16 @@ <h3>Key Requirements</h3>
       <ul>
         <li>Hardware:
           <ul>
-            <li>Raspberry Pi 4 (or similar Arm SBC).</li>
+            <li>Raspberry Pi 5 or 4 (or similar Arm SBC).</li>
             <li>USB microphone.</li>
             <li>Speaker via 3.5 mm jack or HDMI.</li>
+            <li>Where possible, aim to use the CPU without additional accelerators/hats. Solutions that are well-optimised through use of Quantisation, KleidiAI, and appropriate model selection - and therefore able to run entirely on CPU - are of great interest.</li>
           </ul>
         </li>
         <li>Software:
           <ul>
             <li>Python with PyAudio for audio I/O.</li>
-            <li>Coqui STT or fine-tuned wav2vec2 for ASR.</li>
+            <li>Coqui STT or fine-tuned wav2vec2 for ASR, or similar.</li>
             <li>eSpeak-NG or Festival for TTS.</li>
             <li>Custom Python logic for intent recognition and command execution.</li>
           </ul>
@@ -509,7 +515,7 @@ <h3>Objective</h3>
       <h3>Project Description</h3>
       <p>
         Students will choose a lightweight object detector (e.g., MobileNet-SSD, YOLOv5s), convert it to an
-        edge-optimized format (TensorFlow Lite / ONNX Runtime), and integrate it with an OpenCV video pipeline.
+        edge-optimized format (TensorFlow Lite / ONNX Runtime / ExecuTorch), and integrate it with an OpenCV video pipeline.
         Detected anomalies should trigger timestamped logs or saved clips.
       </p>
       <div style="margin:24px 0; display:flex; justify-content:center;">
@@ -526,16 +532,17 @@ <h3>Key Requirements</h3>
       <ul>
         <li>Hardware:
           <ul>
-            <li>Raspberry Pi 4 with proper cooling.</li>
+            <li>Raspberry Pi 5 or 4.</li>
             <li>Raspberry Pi Camera Module v2 or USB webcam.</li>
             <li>High-write-speed microSD card.</li>
+            <li>Where possible, aim to use the CPU without additional accelerators/hats. Solutions that are well-optimised through use of Quantisation, KleidiAI, and appropriate model selection - and therefore able to run entirely on CPU - are of great interest.</li>
           </ul>
         </li>
         <li>Software:
           <ul>
             <li>Raspberry Pi OS.</li>
             <li>Python, OpenCV.</li>
-            <li>TensorFlow Lite / ONNX Runtime with a pre-trained, quantized detection model.</li>
+            <li>TensorFlow Lite / ONNX Runtime / ExecuTorch with a pre-trained, quantized detection model.</li>
           </ul>
         </li>
       </ul>
@@ -576,7 +583,7 @@ <h3>Learning Outcomes</h3>
 
     <h3>Objective</h3>
     <p>
-      Build a fully local, real-time speech-to-speech translation system optimized for Arm-based CPUs, leveraging SME2 where available (preferred) or NEON/NPU otherwise. The system must perform speech recognition, LLM-based translation or semantic rewriting, and speech synthesis entirely on-device, meeting mobile latency, power, and thermal constraints.
+      Build a fully local, real-time speech-to-speech translation system optimized for Arm-based CPUs, leveraging SME2 where available (preferred) or NEON or an onboard NPU otherwise. The system must perform speech recognition, LLM-based translation or semantic rewriting, and speech synthesis entirely on-device, meeting mobile latency, power, and thermal constraints.
     </p>
 
     <h3>Project Description</h3>
@@ -590,7 +597,7 @@ <h3>Project Description</h3>
         <li>LLM-based translation or semantic rewriting, and</li>
         <li>Text-to-speech (TTS) synthesis,</li>
       </ol>
-      to produce natural, fluent spoken output in Language B. All inference must run locally with no cloud dependency, demonstrating efficient use of Arm CPU acceleration and mobile-friendly optimizations.
+      to produce natural, fluent spoken output in Language B. All inference must run locally with no cloud dependency, demonstrating efficient use of Arm CPU acceleration and mobile-friendly optimizations. To see all SME2-enabled devices, and for resources to get started, see this related Arm Developer Labs project. <a href="https://arm-university.github.io/Arm-Developer-Labs/2025/11/27/Edge-AI-On-Mobile.html">Edge AI on Mobile using SME2</a>
     </p>
     <div style="margin:24px 0; display:flex; justify-content:center;">
       <iframe width="560" height="315"
@@ -607,8 +614,10 @@ <h3>Key Requirements</h3>
       <li>Hardware:
         <ul>
           <li>Arm-based smartphone CPU</li>
-          <li>SME2-enabled device preferred; otherwise NEON-optimized CPU or optional onboard NPU</li>
+          <li>SME2-enabled device preferred; otherwise utilise an Arm-based CPU and NEON instructions or optional onboard NPU</li>
           <li>Microphone and audio output (speaker or headphones)</li>
+          <li>Where possible, aim to use the CPU (leveraging SME2 if available) without using an onboard NPU. Solutions that are well-optimised through use of Quantisation, KleidiAI, and appropriate model selection - and therefore able to run entirely on CPU - are of great interest.</li>
+          <li>No cloud inference permitted</li>
         </ul>
       </li>
       <li>Software:
@@ -632,8 +641,6 @@ <h3>Key Requirements</h3>
               <li>Examples: FastSpeech2 + HiFiGAN, VITS-lite</li>
             </ul>
           </li>
-          <li><strong>No cloud inference permitted</strong></li>
-          <li><strong>Quantization and Arm-specific optimizations required (SME2/NEON)</strong></li>
         </ul>
       </li>
     </ul>
@@ -840,4 +847,4 @@ <h2 style="color:#7c3aed; margin:0 0 15px; font-size:24px;">Thank You!</h2>
       });
     }
   });
-</script>
+</script>